Roblox Stock

Good afternoon, everyone. My name is Regina, and I will be your conference operator today. Welcome to Roblox's Second Quarter 2026 Earnings Conference Call. [Operator Instructions] I will now turn the call over to Jaime Morris, Roblox's Head of Investor Relations. Jaime?

Good afternoon, everyone. Thank you for joining us to discuss our Q2 2026 results. With me today is Roblox's Co-Founder and CEO, David Baszucki, and our Chief Financial Officer, Naveen Chopra. Before we begin, I would like to remind you that our commentary today may include forward-looking statements, which are subject to risks, uncertainties and assumptions that could cause actual results to differ materially from those described in our forward-looking statements.

A description of these risks, uncertainties and assumptions are included in our SEC filings, including our most recent reports on Form 10-K and Form 10-Q. You should not rely on our forward-looking statements as predictions of future events and we disclaim any obligation to update these statements, except as required by law.

During this call, we will also discuss certain non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics can be found in our shareholder letter and supplemental materials, which are available on our Investor Relations website.

With that, I'll turn the call over to Dave.

Good afternoon, and thank you for joining us. Today, I'll start with a few remarks about the quarter, and then I'll turn it over to Naveen to discuss results and guidance and then I'll conclude with some important updates about how we plan to achieve our long-term goals.

We remain steadfast in our goal to capture 10% of the global gaming market and an even greater share of the U.S. market. Against the backdrop of the incredible growth we delivered last year, in Q2, we delivered healthy year-over-year growth in users, hours and cash flow. And well bookings [ Gravel ] within our guidance range it was below our internal goal, which we will discuss in detail on today's call. Before I hand it over to Naveen, I want to give some updates on the business.

First, let's talk about safety. In the first part of this year, we made the decision to age check everyone on our platform and to limit chat only to those users who have aged check. We've made significant progress in the quarter. As a result, age check penetration has increased to 57% globally and in the first countries where we rolled it out, Australia is approaching 80%, and the U.S. and U.K. are at 70%.

With the under 18 segments, we're making enormous progress with the U.S. under 18 at 75% and in addition, of course, to our filters, industry-leading safety, we don't allow image sharing in chat and in conjunction with age check, we implemented age-banded communication.

We've also introduced new comms products. We'll share a bit more about that. And on the content side in June, we successfully rolled out age-based Roblox Kids and select accounts. We believe accurate age data amplifies our conviction in the long-term growth potential for the platform.

On the discovery side, in the last few months, we shared with our community our strategic decision to focus our discovery algorithms directly on measured long-term retention. We have seen this impact monetization primarily in the U.S. under 13 cohort and this has been exemplified by a lower frequency of impressions with games that emphasize short-term monetization rather than long-term retention.

It's still early, but our data shows a positive lift in quality, stickier engagement games and the benefits of higher retention, we believe, outweigh the near-term impacts we're seeing as improved retention and engagement compound into future growth.

I'm going to now turn it over to Naveen to address some specifics and then I'll be back to talk about our future plans.

Thanks, Dave. I'm going to share a few details about the quarter and then also discuss updates to our outlook. For the quarter, revenue grew 36% to $1.5 billion and bookings grew 8% to $1.6 billion. As Dave mentioned, this was at the low end of our guidance, and we'll discuss more about that in a moment. We finished the quarter with 123 million DAUs, up 10% year-over-year and hours grew 5% to $29 billion.

In addition to building on the existing scale we have in the United States, we continue to see strong year-over-year growth in many international markets. For example, DAUs in Japan grew 67% and DAUs in India grew 64%. And similarly, ours growth in Japan and India was 61% and 59%, respectively.

We also saw users in Russia come back online as we were unblocked there late in the quarter. On the bottom line, operating cash flow was up 60% year-on-year to $318 million and free cash flow of $294 million was up 66% year-over-year.

From a content perspective, our platform is getting more diversified, while maintaining a steady velocity of new content. Case in point, 3 years ago, our top 10 games accounted for about 30% of all the hours engaged on Roblox. Today, the top 10 games account for approximately 20% of ours.

And in terms of user metrics, new user sign-up activity improved throughout the quarter, benefiting from seasonality and the reinstatement of Roblox in Russia. Retention of existing users remain stable and engagement hours were in line with our expectations.

Monetization, however, as measured in bookings per hour was below our forecast, particularly with under 13 cohorts and was the primary factor impacting our bookings performance. We attribute the unforeseen monetization shortfall to a greater-than-expected shift of engagement from high monetizing 2025 vintage viral games to a combination of new and evergreen experiences with lower hourly monetization.

This underlying mix shift was compounded by changes in our recommendation algorithm, which optimizes for long-term retention and is, therefore, providing more impressions for highly retentive games at the expense of near-term monetization.

And in Q2, the bookings impact of this trade-off on younger users was greater than we anticipated. Although these discovery changes add to the variability of our business, our testing suggests that the long-term benefit to retention will more than offset the short-term monetization hit. Particularly as we continue to improve the algorithms by doing things like using ACC data to better target recommendations by region and age cohort, which, over time, should help mitigate the monetization headwind we saw in Q2.

Another albeit less significant factor in the monetization shortfall was the impact of our decision to disable the sale of cross-experience game passes. Looking ahead to Q3, we expect sequential DAU trends to benefit from seasonality, the unblocking of Roblox in Russia and our discovery changes.

But monetization weakness is likely to continue. And as a result, we forecast Q3 bookings to be between $1.58 billion and $1.65 billion, which, given the tough comparisons translates to a year-over-year decline of 14% to 18%. From a margin perspective, the reduction in bookings will result in fixed cost deleveraging.

Additionally, investments in AI-powered initiatives like build, Roblox reality and moments are expected to result in higher infrastructure costs. For Q4 increasing variability and continued updates to our platform lead to a wide range of potential outcomes. And therefore, we are not providing revised full year guidance at this time.

While our expectations for the remainder of the year have changed significantly, we have conviction that we're making the right trade-offs to continue our role as an industry disruptor. Investments in AI, content diversification, long-term retention and safety, though creating near-term friction position us to maximize our share of the global gaming market.

With that, I'll pass it back to Dave to talk about some of our future plans.

Thanks, Naveen. Since we started the company, we've been on a mission to support play around the world, and we believe the world needs more play. We can all see it right now given in light of really everything going on in the world. And this fundamental need is part of what gives really all of us conviction in Roblox's ability to deliver long-term growth in excess of 20%.

We set our goal of expanding play everywhere with a targeted 10% of global gaming running on Roblox. And in the U.S., we've shared that our goal is to go well beyond 10%. Today, we estimate that 20% of USA under gaming revenue flows through Roblox. We've done this through the power of our platform and the creativity and diversity of user-generated content and in this cohort, we validated the power of our ability to unify discovery, economy, tools, engines, safety and cloud infra.

While we did experience softness in our younger users monetization this quarter, we believe our monetization system is fundamentally strong and will compound over time with additional levers, including subscription and advertising revenues.

We've established unlike really any other company at our scale, our commitment to setting the global standard for safe digital engagement and with younger users, in particular, long term, we believe this is a tremendous benefit to have such strength among younger cohorts, and this will contribute, we believe, to fuel long-term growth.

The 18 and over market is 80% of the $200 billion global gaming market, and we believe we're uniquely poised to expand play in this 1 and over market in the same way we have for under 18. Today, in the U.S., our '18 and up cohort represents nearly 1/3 of our now age checked daily active users and our U.S. 18 to 34 cohort daily active users and hours are growing at approximately 40% year-on-year.

And our 13 up cohorts around the world continued to show strong growth. In addition to our viral growth worldwide in under 18, we have the opportunity to power 18 and up growth with increasing paid acquisition as we continue to improve product market fit.

And besides safety, the majority of our product and technical investments supporting growth in 18 and up also benefit younger audiences. We've recently launched 3 initiatives that will support our continued growth in the 18 and up market segment.

First, as we shared, we're removing self-imposed limits on the types of games we support on Roblox. Second, we've redefined our app to align with physical play including discovery. And third, we're reconnecting with our core promise that anyone can be a builder.

First off, we want all types of gaming on Roblox. We're not that far away from having the technology in place to support the vast majority of games that people play online. On one end of the spectrum, we believe players are looking for multiplayer experiences that will ultimately mirror physical reality.

Even since our start, we've had the goal of photorealism and unlimited multiplayer scale and with our announcement of Roblox reality, we're well on our way to being the first company to bring photorealistic multiplayer to life and to the masses. We see a future where games really will become indistinguishable from movies for certain types of games.

And we can see this quality today in offline AI video tools we believe we'll soon see this in real time with robots reality. Now gaming is also simultaneously spontaneous and single player and 2D for many people and up until today, we've been focused on 3D, and we've held back on initiatives to support high-performance 2D tooling and gameplay.

Going forward, we're doubling down on our support for everything from 2D puzzle games to strategy. With Roblox Reality and a focus on 2D performance, we're expanding the bookends of what we believe will be possible on Roblox. This week, we've announced a unified vision of our app that includes bringing video discovery and creation front and center to all users.

We've announced this week, we're integrating moments into our homepage for 16 and up users, and we've begun testing for 16 users in Canada, New Zealand and Singapore. We're expanding the moment feed to include short-form influencer video in addition to game captures, and we believe, ultimately, moments will be available for all with appropriate and safe educational content for younger players.

Ultimately, we'll expand moments with tiles that let users watch creators or view live gameplay in real time, so you can see what your friends are playing. Our early tests show substantial gains in moments user retention, and we're very optimistic about moments.

We'll ultimately be moving to full 3D games in the moments tab, and this will become integrated with games created through build. Together with build, this means you can watch a game in action and then play it with one tap creating really a unique path all the way from discovery to this new wave of content from build.

On the communications side, continuing our dialogue from last quarter, we're committed to leveraging our world-class industry-leading safety infra to build the highest quality integrated voice, video and text comms on the Roblox platform. We've already shipped global chat and quick words.

And today, a significant portion of our users still do utilize off-platform communication tools. our road maps to provide users the option to video chat with friends or stream with their Roblox Avatar.

Because all comms that originate within Roblox will company users as they move between games, they deliver a safer, richer, more immersive connection than any off-platform alternative.

And as a first step in Q3 will be deploying inexperienced friend chat, which brings real-time text messaging directly into any robots game. Creating a game is much more complicated than writing code. Games are built with code, images, graphics, audio tracks, 3D sees, avatars and of course, with enormous creativity --we believe we have a unique opportunity to accelerate game creation based on integrated and vertical stack.

Last week, we announced Build, an extension of our vision that everyone in the world can be a game builder and we're leveraging our internally generative AI models to create a conversational interface. So anyone who can imagine something can describe it and shape it with others and then ultimately bring it to a playable experience to life and minutes.

For the creators that are already building with us -- we believe this will be about practical leverage, helping them build faster and to reach the right audience more effectively.

Build will be fully integrated with Roblox Studio and the AI powering build will allow users to start an experience and build and ultimately finish it in Roblox studio.

We're already live in New Zealand with Build. I just want to highlight we're seeing a large number of creations flowing through the system already, Literally, people fulfilling our vision of you make the game.

Now just as coding, it's not just the model, but we believe with Build, it's the integration of a harness, tools and skills. And ultimately, our vision is many models unified in 1 harness using our cloud back end in infra to come full circle. The other cool thing about build is it really integrates with what we said about removing the bookends.

Many people who are using Build will want to build 2D experiences or 3D and build will leverage this with our back end to support everything in between. Given our discovery focus on long-term retention, we're comfortable that we're in a position to appropriately surface the high-quality creations with build even as the quality of creation increases.

We're really navigating a period of normalization, we believe, following a year of massive growth. We have tremendous scale and engagement, which presents us with a unique opportunity to amplify our long-term potential. To capitalize on this momentum, we're doubling down on our vision that the world needs more play.

And we're bringing to life really the original vision of our company which is you make the game by expanding our global audience, broadening our content offering and deepening engagement for our users and creators. I've never been more excited about our mission and more excited and confident in the road ahead.

With that, we'll pass back to the operator for questions.

[Operator Instructions] Our first question will come from the line of Matthew Cost with Morgan Stanley.

I guess you've done a really helpful job on this call laying out what's the long-term vision of what you're trying to build with the algorithm changes and the Kids and select accounts you've done a great job of explaining what that's meant to get to over the long term.

I guess between here and there, what are you watching? And what should investors be watching to help determine when the engagement headwinds from those changes have peaked? And then sort of connected to that, when we think about the expectation of user growth sequentially in the third quarter, is that just a function of seasonality?

Or are there any improvements in the trajectory as you've exited the second quarter?

I have a list, Naveen, do you have a listed? Do you want me to go first? Or do you want to go first?

You can go first, and I'll catch the last couple of parts to that question.

Yes. Matthew, great question. I think a couple of things to watch. In our U1 segment, we've laid out on this call really where we're going, and that includes we believe, finally, we'll be fulfilling that original vision of you make the game. Not every user can figure out how to use studio or make a game.

What we see already in New Zealand is more people per day on build than on studio. And we do think, ultimately, AI creation of gaming is going to be something that sits side-by-side playing games.

So I think watch for frequency in under 18, we also expect to be fairly generous in the amount of AI or tokens in the build process. But for those users that are using build all day long, we will, within our Roblox Plus subscription, add the ability to increase will be called token somewhere else. We'll have our own name for it. watch our subscription for under 18 as well.

And then I would say also as we roll out continued comms platform watch for that, over 18, we're laser-focused on a couple of cohorts in that and the growth rate in those cohorts.

And we're laser-focused on the type of content that's being created by our creator ecosystem that we believe will be more and more retentive in those cohorts. So I would watch the growth rates in over 18. I shared a bit of what they are today because those point to 80% of the market that we can -- we believe we can do the same thing and as we've done under 18.

And then just -- just a couple of the things you asked about, Matt, in terms of sort of the near-term friction and then DAU trends I just would clarify, you talked about near-term engagement friction. That's really not what we're seeing. We're actually pretty happy with the engagement trends that we've seen.

I talked about some of the health that we've seen in sign-ups, retention, et cetera. So engagement looks good. The weakness that we saw that was inconsistent with what we expected when we put our original plans together for the quarter was really around monetization.

And that is going to continue as we continue to evolve some of what we're doing around discovery and recommendations. The DAU trends themselves, seasonality is definitely a big factor there. We're going to get a full quarter benefit from having Russia back online and so those are probably the 2 big ones that I would highlight with respect to DAU trends.

Our next question will come from the line of Ken Gawrelski with Wells Fargo.

Two, if I may, please. Could you maybe first touch upon the updates to the Discovery engine? You did a good job of explaining that in the letter, but maybe if you could talk about when you expect to see some of these longer-term -- the longer-term retention benefits filtering into both engagement gains and also potentially monetization gains. Just if you could walk us through kind of maybe the time line and your expectations there.

And then second, if I may, please, Naveen, could you just talk about as we're going through this transition and the push for over '18 and new content. Could you talk to us about any changes you might expect in the cost structure?

How either from either a Devex standpoint, from a tech standpoint or platform standpoint, how might the financial profile of the business look in success 12 to 24 months out relative to the prior 12 to 18 months.

I'll start with the discovery algorithms. I touched on as we continue to refine discovery, what any discovery algorithm is really trying to do is connect users with experiences that both drive long-term enterprise value and simultaneously give signals to creators of what types of experiences are rewarded with discovery.

So it's a complicated process. It involves both feedback to creators as well as feedback to users. And watching social if we've implemented and improved our algorithms over the last few months, there's been a resounding support in the feeling that moving more and more to evergreen type games, games that keep players around for a long-term essentially the exact opposite of what might be called not that we're saying we had it, but more click baby or kind of cash grabby type games.

We want to be as far in the evergreen segment as we can. We made the decision as we started going down this next gen to go longer into the time frame to measure signals longer.

We made the decision to use that feedback loop to update the algorithm even as we're running additional tests. And we've really churned our discovery system from something that's baked in to something that in a way is constantly self-improving and self estimating our best guess at what user game pairs will drive the best long-term retention.

We're balancing that more and more with what we think is the optimal mix of long-term retention and long-term monetization -- and in our experiments, as Naveen said, we can see shorter-term reductions in monetization, but then crossover points and longer-term improvements. We're not going to share exactly what the crossover point is, but this is the reason we implemented and have gone with this.

So it's directly measured retention. It's less projecting. It's more direct measure of both retention and monetization. And one other thing is we introduced build and we expect the volume of creations to grow much higher on Roblox, we believe this algorithm is very resilient to both finding great new build games that will be playable by all and simultaneously not flooding users with games that might have an AI kind of name to them that aren't as long-term retentive.

All right. And then, Ken, with respect to your question on the cost structure. There's a couple of important things to highlight here. First, our view of the long-term margin potential of the business really has not changed, meaning the kinds of things that are going to drive margin expansion over time, are still in place.

Obviously, continued bookings growth generates fixed cost leverage we continue to expect to see mix to lower-cost platforms that will help COGS and Devex increases are something that we do intend to continue to pursue but those ultimately are a function of us realizing fixed cost leverage in other parts of the business.

So that's point one. Point 2 in, call it, the near to medium term, the investments that we're making in a lot of AI-powered product enhancements, I do expect to result in some elevated levels of infrastructure expense. But it's important to realize -- but unlike our inter expense historically, this is more of a fixed cost because historically, infrastructure investments have been a function of hours and users on our platform.

These infrastructure costs are largely being driven by model training. And so we're moving aggressively in that direction. But over time, we expect, given the sort of more fixed cost nature of that as bookings grow, we'll start to see leverage on those infrastructure investments as well. But as you saw in the Q3 guidance, there is going to be some pressure in the short to medium term as a result of the incremental cloud GPU capacity that we need there.

Our next question will come from the line of Eric Sheridan with Goldman Sachs.

I appreciate all the comments about the progress you're making towards being 10% of the gaming market or maybe even more in North America. How do you think about the competitive landscape continue to evolve both within gaming but also among younger users where they're sort of conflicting narratives about being online versus off-line gaming versus non-gaming in terms of time spent and how you think continually about positioning yourself relative to the gaming landscape and even a more broader definition for the competition for time and engagement.

I really think the broader definition is a bigger thing than Roblox, and it is our vision that the world needs more play and play is something that people do with each other. It's something that involves not just playing but creating it involves being with others, communicating with others being yourself and watching them.

And really as we've redesigned our app and brought the build as well as moments to the homepage, it's really part of that vision of mirroring the physical world with what's happening in our digital world of supporting play.

We think -- we think just like in the real world, where play involved sometimes creating the game, I'm optimistic that we're going to see a volume of creation with build unlike any volume we've ever seen with Roblox Studio while at the same time pro devs and devs making extremely complex properties will utilize all of the capabilities of build as they move to large teams and studio to build much more complex type creation.

So I do think what is gaming is changing. I think we're going to enter an era just like 10 years ago where most of us didn't know how to edit video. And now all young people, essentially, many of them know how to edit video on their phone.

The same thing, I think, is going to happen in gaming where really with AI acceleration, everything from 2D puzzle games to complex 3D multiplayer games are going to be created on mobile, enhanced in roadblock studio.

And that's going to really change the whole landscape, I believe, of what gaming is. So I do think the use of today are changing, I think, play as universal and I believe we're moving towards that original vision of really everyone on our platform is a creator and a builder.

Our next question will come from the line of Chris Schoell with UBS.

Great. Naveen, we've seen you announce the buyback since the last earnings call. Just as free cash flow for the business scales in the coming years, can you remind us how you think about capital allocation and how you'll prioritize buybacks with organic investment in M&A? And what impact does any -- if any, does the stock performance have on your willingness to maybe lean in further on the buyback relative to alternative uses of capital?

Yes, sure. I think it's -- we're pretty clear with respect to our capital allocation strategy, meaning the first and foremost for us is making sure that we have the ability to aggressively invest in organic growth.

That largely takes the form of R&D, people, tokens, infrastructure, things of that nature, we try to be very diligent and very efficient with all of those investments, but we also recognize we live in a very rapidly changing world right now.

And we're going to keep a lot of dry powder to make sure that we can be highly nimble with respect to organic investment. That being said, when we look at our balance sheet and our free cash flow trends, we have a lot of firepower to do both what I just said with respect to organic investment, but also to support a buyback and potentially M&A.

Our M&A historically has been largely focused on acquiring technical talent, and we have ample ability to continue to do that but also potentially bigger things in the future to the extent there's strategic or industrial logic to it. In terms of stock price performance affecting how we think about the buyback, 2 comments on that.

One, the way we have structured our buyback, which is largely to offset dilution from employee equity grants thus result in us effectively buying more stock when the stock price goes down and buying less when the stock price goes up.

So there is some sort of natural adjustment built into the mechanism. We do have the ability to accelerate that or supplement it, and those are obviously things that we'll continue to evaluate.

Our next question will come from the line of Omar Dessouky with Bank of America.

I think you guys explained clearly why you gave your third quarter guide. But I was hoping that you could give us perhaps some guardrails around the fourth quarter and full year, given that you're not guiding the full year anymore.

Number one is '26 still a growth year. Would you be comfortable telling investors that 2026 bookings will be higher than 2025 bookings in total? And when we think about those factors that caused you to give your third quarter guide, how is the fourth quarter -- how are those factors going to improve or trend in the fourth quarter.

What's getting better in the fourth quarter versus the third quarter? Just to kind of help us put some guardrails around what the fourth quarter could look like.

Yes, I'll take that, Omar. As I said, we're not going to provide guidance for the fourth quarter right now simply because of all of the moving pieces that I think we articulated both in our letter and some of our comments. We feel good about where things are trending in the long term.

We are navigating, as Dave said, both normalization from last year and also a number of platform changes across discovery, safety and now launching some very meaningful new products with things like build and moments. And the reality is it's very early days for many of those new initiatives.

There's no model that we can look at that's going to say is exactly what they're going to do over the next 3 to 6 months. And we want to make sure that we're moving as quickly as we can on all of those initiatives. So you have to assume that the things we do in Q3 and Q4 are going to be consistent with what we said we want to accomplish over the long term, but we can't give you any more specific guidance than what I've shared.

Our next question will come from the line of Clark Lampen with BTIG.

Maybe, Naveen, sort of the first and 2 questions, if I may, would be in the shareholder letter, you mentioned you were comfortable based on the internal testing that longer retention would ultimately overcome a reduction in hourly monetization.

Could you elaborate a little bit more on that or maybe what you had seen when you had run those sort of comparisons? Second one is very quick, just on Morpheus, could you help us think about either the magnitude of financial commitment or potentially help size the IPS headwind that we might see this year?

Yes. So on the first one, I think Dave talked about this a little bit, too. So maybe just to put a little bit of a finer point on it. We've been running a lot of experiments and continuing to evolve these algorithms.

And what we have seen when we sort of AB test them relative to people on, call it, older discovery algorithm, is that very quickly, there is a change in user retention, which ultimately that's a compounding effect for our business because ultimately, more users is what drives the robot flywheel that also shows up relatively quickly in incremental hours of engagement.

However, there is a pretty immediate hit to dollars per hour. That's sort of the monetization metric that we look at. And we do believe that based on the curves that we've seen with, call it, a few weeks of data for each of these experiments, that those lines cross, meaning the benefit of the incremental retention will more than outweigh the short-term hit to bookings.

We're not going to put a specific time line on that right now, in part because we're continuing to evolve those algorithms. And our focus right now is actually on finding ways to improve the algorithms such that we can reduce the near-term bookings impact while retaining the retention benefit. And there are some promising irons in the fire on that, but too early to say exactly what the impact will be.

Go ahead, David.

I was just going to add the longer out we look at signals and directly measure them. The longer we pick up not just early revenue but longer revenue in conjunction with that retention, we are trying to optimize a blend of long-term retention and long-term monetization that's directly measured as opposed to overly extrapolating from short-term monetization signals that may be chatty and potter out. .

And then, Clark, with respect to the second part of your question, on the magnitude of the in for investments and sort of the headwind of that great. I think you asked the question in the context of Morpheus is one small part of many different AI initiatives that we have going on. So I'm going to answer the question sort of bigger picture.

And probably the simplest way to think about it is if you look at our margin guidance for Q3 and you look at the margin compression relative to the year ago period. I would say that about half of that is just fixed cost deleveraging coming off the back of the bookings reduction and the other half is related to these investments in AI.

And those are the things that power new features like Build and moments, a lot of the improvements we're making on safety as well as training models for things like robots reality. So all really important, exciting stuff that requires incremental investment.

And I want to highlight with Morpheus the significance of what we're working towards. Our goal is to provide photorealistic multiplayer gaming which does not exist in the world today. On one end of the spectrum, we're starting to have off-line video models that start to approach a 4K movie, on the other end, 3D gaming technology is getting better and better and better, but it's not quite photo realistic.

We believe a hybrid approach, which is both 3D synchronization in the cloud with and kind of user super up sampling to get to photorealistic is the ultimate way to do this. And the cost of infra here will initially be offset by subscription. Games will run both in regular Roblox mode as well as in super up sample Morpheus video photo reality mode, and we will initially expect to charge subscription to access that.

Our next question will come from the line of Clay Griffin with MoffettNathanson.

Dave, I'm curious to get your thoughts on how you see the 2D content that you're enabling with some of the announcements that you've made today. And typically, we'd like to think of mobile as more of a single player experience in many cases.

So do you envision 2D Mobile content being sort of an acquisition funnel for players that are outside of your platform today? Or do you think that there's demand for that type of content on the platform today?

And how do you think about what might make 2D continental Roblox differentiated from the typical to the mobile experience today?

Yes. I think when we take a step back and go to that original vision of you make the game, there's a really good conjunction, we believe, between what we're doing with Build and with what we say about removing the bookends of the type of content we want on our platform.

When we start putting Build in front of people, they don't constrain themselves to saying, I want to build a 3D multiplayer [indiscernible] course or I want to build a puzzle game where you do this or I want to build a multiplayer obstacle course game.

So there's really a great intersection in supporting all that people want to build on build. The other thing is, I think many people don't differentiate necessarily between a 2D single player game, a 2D multiplayer game, a 2.5 orthographic game a 3D multiplayer game they just think of play.

And as we get into more and more older cohorts, some of those cohorts have an ample appetite for 2D but we believe huge advantages in any experience that's built on Roblox, runs in multiple languages, runs around the world is socially enabled is backed by our economy is backed by our infra and can intermingle throughout them.

So we see this really as just approaching having a wide range of gaming on the platform. And ultimately, it works with build and that we're not going to limit what people can build when they go and imagine creating a game and showing it to their friends.

Our next question comes from the line of Andrew Marok with Raymond James.

Maybe one, can you please talk a bit about the differences in monetization patterns between under 13 players and other cohorts that may have driven an outsized impact this quarter.

Was it specifically related to the viral hits from last year or is it a general rule that these younger users are maybe more volatile in their monetization patterns?

Yes, I'll take that. I think it's more the former, meaning if you just look at sort of the percentage of hours consumed on the platform last year that was coming from these very large and high monetizing viral hits it was very concentrated.

And those games did appeal, in particular, to younger users. And so when we have seen this mix shift from -- with those games getting replaced by, call it, games with more normal levels of monetization that shows up most significantly with these younger users. I don't think there is anything sort of structure with respect to changes in younger user behavior.

It's really just the games that are popular at this point in time, don't have that sort of very elevated level of monetization that we saw last year.

Our final question will come from the line of Aaron Lee with Macquarie.

Appreciate it. Maybe just to touch on incubator and the 18+ initiatives. It seems like that's progressing nicely. Can you just talk about like what you've learned through this process so far? And maybe how that informs how this program could evolve going forward? Or any other initiatives in the pipeline to really drive this business forward?

Yes, I'll comment on it. We have a lot of content coming not just from incubators, but also studio partners and our existing community. We do see ultimately as a UGC platform, the power of that -- but we do also see the ability almost from a sales engineering or an expansion of how we see DevRel to be very intimate with the top creators on our platform and to be very intimate with them as we help guide them to what performs well on our platform.

We really have a focus on a lot of the tech we've recently introduced with our incubators. We've introduced the ability for much higher performance avatars for much higher performance worlds that run very well on low end to Android as well as on high-end PC.

We've introduced a compositing technology called Slim that allows very complicated avatars to run at high performance on phones. So all of this working with these partners, we believe, is a huge technical opportunity for them to build games for us and of course, they're also working on our DevEx premium of 50% for games consumed over 18.

So really, it's not just the incubator. It's many of our corporate studios. We're working with as well as our existing developer community, and we're seeing great quality coming out of them.

That concludes the question-and-answer session. And I'll hand the call back over to Dave Baszucki for closing comments.

Once again, thank you for all the great questions. It really is an exciting time for us as we move towards fulfilling that original vision we had when we started the company, if you make the game and really for everyone around within Roblox helping to support more play everywhere around the world. Thank you again for your questions and attention.

This concludes today's call. Thank you for joining. You may now disconnect.

Hello, and welcome to the Roblox Corporation Annual Meeting of Stockholders. Please note that this meeting is being recorded. [Operator Instructions] I'd now like to turn the conference over to the company.

Good morning. I'm David Baszucki, Chair of the Board of Roblox Corporation, and it is a pleasure to welcome you to Roblox 2026 Annual Meeting of Stockholders. I will act as Chair of this meeting. I've asked Mark Reinstra, our Chief Legal Officer and Corporate Secretary, to serve as Secretary for this meeting, record the minutes and conduct the formal portion of the meeting, including the opening and closing of the polls on all matters to be voted at this meeting.

We're holding our annual share -- Annual Stockholders' Meeting virtually this year as authorized by the Board and permitted under Nevada law and Roblox's bylaws. Participation through the meeting platform constitutes presence in person at the meeting for purposes of Nevada law. Please note, this meeting is being recorded. We will conduct the business portion of our meeting first and answer questions at the end of the meeting. Please note, we will not be making any presentation on the business or financial condition of the company at this meeting. It is now shortly after 8:00 a.m. Pacific Time on May 28, and this meeting is officially called to order.

Here with me today are 6 members of our Board of Directors: Gregory Baszucki, Christopher Carvalho, Dennis Durkin, Jason Kilar, Anthony Lee and Andrea Wong. Also attending this meeting are representatives from Deloitte, our independent auditor, Dan Lee and [ Kaitlyn Lieberman ].

Now over to Mark to conduct the formal portion of the meeting.

Thank you, Dave. The agenda and rules of conduct are accessible by clicking the Documents tab on the top right of your screen. In order to ensure the business of the meeting proceeds efficiently, we ask that you observe those rules. The annual meeting is being held in accordance with Roblox articles of incorporation, bylaws and Nevada law.

During the formal meeting, we will address the matters described in the definitive proxy statement filed with the Securities and Exchange Commission on April 16, 2026. I have an affidavit confirming that the notice of meeting, proxy statement and proxy card were mailed and made available on April 16 to stockholders of record as of the April 1 record date.

These records will be filed with the minutes. I have been appointed to act as Inspector of Elections. I have signed an oath to carry out my duties with strict impartiality in accordance with the best of my ability, which will be filed with the minutes of the meeting. As required by Roblox's bylaws as Inspector of Election, I will ascertain the number of shares outstanding and the voting power of each and determine the shares represented at the meeting the validity of proxies and ballots, the count of all votes and ballots and the final voting results, and I will certify those determinations for the meeting records.

We have present in person, including those participating remotely or by proxy, holders of a majority of the voting power of shares issued and outstanding and entitled to vote at this meeting, which constitutes a quorum. Therefore, the meeting is duly constituted, and we may proceed with the business of the meeting. We will vote by proxy and in person electronically. The meeting platform is designed to verify the identity of each participating stockholder and proxy holder and to give stockholders a reasonable opportunity to participate in the meeting, vote on the proposals, communicate during the meeting, and read or hear the proceedings as they occur.

Each holder of Class A common stock is entitled to 1 vote for each share held of record at the close of business on the record date. Each holder of Class B common stock is entitled to 20 votes for each share held of record at the close of business on the record date. If you've already voted, no further action is needed. If you would like to vote now or change your vote, please click on the Vote My Shares tab on the right, top right of the screen and follow the instructions there. Stockholders who wish to submit a question during the meeting should click on the questions tab at the top right of the screen. Type your question into the text box and then click the submit button.

Today's votes will be combined with those already received, and I will share preliminary results before we adjourn. It is 8:05 Pacific Time on May 28, 2026, and the polls for each matter to be voted on at this meeting are now open.

The first item of business is the election of directors. The company's Board presently has 8 members and is divided into 3 classes, each with a 3-year term. There are 3 Class I directors, 3 Class II directors and 2 Class III directors. As detailed in the proxy statement, the Board has nominated David Baszucki, Gregory Baszucki and Dennis Durkin to serve as Class II directors, all of whom are currently serving as directors because no additional nominations were submitted by stockholders in accordance with our bylaws, I declare the nominations closed.

Election of each director requires a plurality of the votes cast, meaning the nominees who received the largest number of votes cast for their election will be elected. Withheld votes, abstentions and broker nonvotes are not considered votes cast for a nominee and have no effect on the outcome. The Board recommends a vote for each nominee and all proxies solicited by the Board will be voted in accordance with the instructions of the stockholder, provided that any proxies that were dated and signed without giving specific voting instructions will be voted in accordance with the Board's recommendations. If you are voting on this proposal today, please ensure your vote is submitted now.

The next item of business is an advisory vote to approve the compensation of our named executive officers. Approval of this proposal requires that the number of votes cast for the proposal exceed the number of votes cast against the proposal. Abstentions and broker nonvotes are not considered votes cast and have no effect on the outcome. The Board recommends a vote for this proposal and all proxies solicited by the Board will be voted in accordance with the instructions of the stockholder, provided that any proxies that were dated and signed without giving specific voting instructions for this proposal will be voted in accordance with the Board's recommendation. If you are voting on this proposal today, please ensure your vote is submitted now.

The next item of business is ratification of Deloitte & Touche LLP as our independent registered public accounting firm for the fiscal year ending December 31, 2026. Approval of this proposal requires that the number of votes cast for the proposal exceed the number of votes cast against the proposal. As abstentions and broker nonvotes are not considered votes cast and have no effect on the outcome. The Board recommends a vote for this proposal and all proxies solicited by the Board will be voted in accordance with the instructions of the stockholder provide that any proxies that were dated and signed without giving specific voting instructions for this proposal will be voted in accordance with the Board's recommendation. If you're voting on this proposal today, please ensure your vote is submitted now. I have confirmed there are no questions regarding the proposals.

It is now 8:09 Pacific Time on May 28, 2026, and the polls for each matter to be voted on at this meeting are now closed. No additional ballots, proxies or votes and no changes or revocations will be accepted. Based upon all the proxies and ballots received prior to commencement of the meeting, and subject to final adjustment of the numbers for any votes and proxies submitted here today, I can tell you that. With regard to Proposal 1, the 3 Class II nominees, David Baszucki, Gregory Baszucki, and Dennis Durkin received the largest number of votes cast for such nominees. Each of them has therefore been reelected as a director of the company to hold office until the 2029 Annual Meeting of Stockholders.

With regard to Proposal 2, the advisory vote for the compensation of our named executive officers has been approved because the number of votes cast for the proposal exceeded the number of cast -- votes cast against the proposal. With regard to Proposal 3, the ratification of the independent registered public accounting firm for fiscal 2026 has been approved because the number of votes cast for the proposal exceeded the number of votes cast against the proposal. The final results of voting will be reported on a Form 8-K to be filed with the Securities and Exchange Commission within 4 days of the meeting.

As inspector of elections, I will execute a certificate setting forth the results of voting which shall be filed with the minutes of this meeting. There being no further business to come before the formal portion of the meeting. It is now adjourned.

Now we would like to open things up with stockholder questions and comments. [Operator Instructions]

Since there are no questions have been submitted, I will turn the meeting back over to Dave.

Thank you all for attending today's meeting and for the interest you have shown in Roblox Corporation. We very much appreciate your attendance. And as always, thank you for your support.

This concludes the 2026 Annual Meeting of Stockholders of Roblox Corporation. You may now disconnect. Thank you.

Good afternoon, everyone. This is Kate, and I will be your conference coordinator today. Welcome to Roblox Q1 2026 Earnings Conference Call. [Operator Instructions] Now I will turn the call over to Jamie Morris, Roblox Head of Investor Relations, Jamie.

Good afternoon, everyone. Thank you for joining us to discuss our Q1 2026 results. With me today is Roblox Co-Founder and CEO, David Baszucki, and our Chief Financial Officer, Naveen pro. Before we begin, I would like to remind you that our commentary today may include forward-looking statements, which are subject to risks, uncertainties and assumptions that could cause actual results to differ materially from those described in our forward-looking statements. A description of these risks, uncertainties and assumptions are included in our SEC filings, including in our most recent reports on Form 10-K and Form 10-Q.

You should not rely on our forward-looking statements as predictions of future events, and we disclaim any obligation to update these statements, except as required by law. During this call, we will also discuss certain non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics can be found in our shareholder letter and supplemental slides, which are available on our Investor Relations website. With that, I'll turn the call over to Dave.

Thank you. Good afternoon, and thank you for joining us today. We continue to make progress towards our target of capturing 10% of the global gaming content market on our platform and an even greater share of the U.S. market. In Q1, we had revenue of $1.4 billion, which grew 39% year-over-year, bookings of $1.7 billion, which grew 43% year-over-year. And I'll note that's roughly twice we've shared with investors as our long-term growth trajectory. We generated $629 million in operating cash flow and $596 million in free cash flow, up 42% and 40% year-over-year, respectively. Monthly unique payers increased to $31 million, up 52% compared to a year ago. And we have strong payer growth in international markets and continued solid growth in the U.S. and Canada, up 19% year-over-year on a larger payer base.

We also saw year-over-year user and engagement growth DAUs of [indiscernible] million, up 35% year-over-year and DAUs outside of the U.S. and Canada grew 40%. DAUs in the U.S. and Canada grew 17%. DAUs in Japan grew 96% year-on-year in Q1 and DAUs in India grew 84%. Hours engaged were at 31 billion, up 43% year-over-year. Outside of U.S. and Canada, that number grew 50% and hours in the U.S. and Canada grew 21%, hours in Japan grew 101% year-on-year in Q1 and hours in India grew 91% year-on-year. For our 18 and up numbers, as of Q1, over 18 users represent 26% of DAUs who have age checked, in the U.S., DAUs and hours for the 18 cohort grew over 40% and within that, our 18 through 34 cohort grew over 50% faster than any other age cohort. Additionally, in the U.S., the over 18 and over users on Roblox monetizes over 50% higher than our under 18 users.

As we expected, DAU growth, while very healthy at 35% has declined from the roughly 70% growth rates we saw in the past 2 quarters. User acquisition and engagement was also impacted by our global rollout of age checks to access chat in January. At Roblox, we're committed to setting the global standard for healthy, safe and age appropriate digital engagement, and we're building a platform for all ages as part of the core vision to connect 1 billion users with optimism and civility. As part of this commitment in Q1, we became the first large online gaming platform to introduce age checks to access chat on a global basis. And I want to also note, we're really the first large platform other major announcement in our plans to introduce age-based accounts which leverage our a check technology, and we expect this to roll out globally in June.

I want to note, because we globally introduced age check, we've been able to introduce kids accounts within our core app. These proactive measures are setting a new industry benchmark and also, we've been incorporating really input from policy makers and regulators around the world. Note that not all of our users have age checked even as the percentages continue to grow in the United States, we're at 65% age check. In Australia, where we started a bit earlier, we're at 70% age check. This is always in addition to our robust text filtering technology that we're continuously improving and also our open source of voice safety tech.

We're now better able to understand the impact in the short term of age checking. For communication engagement, we have had a follow-on in edition in the percent of users communicating on our platform just because people right now who have not aged checked, we don't allow them to communicate. In addition, along with age checking, we've now banded communications, so we no longer allow adults to communicate with users under 16 even with our existing industry-leading filters and with no image share. We believe this reduction in [ cans ] does affect both people who have aged checked as well as those who have not because those who have age do have fewer people to communicate with.

Also, as we push towards 10% of gaming, we believe our discovery algorithms should be focused primarily on driving incremental long-term platform retention over short-term monetization, especially to grow our 18 and up user base. We're implementing this transition now. And as we continue to adjust to local customs and regulations, we do foresee some restrictions of content to both non age check users relative to their age range and also as part of region-specific content guidelines. We believe as a result of age check also and reduced communication discovery that we've weighted more towards monetization. We have seen a reduction in app store ratings and we believe this may be contributing to a reduction in organic sign-ups that typically flow from app stores.

We believe the strategic upside of everything we're doing is significant and the right thing to do for the long-term health of the platform. Kids accounts and select accounts offer the long-term opportunity of increasing safety and stability, which in turn drives organic engagement growth in line with our mission to connect 1 billion users with optimism and stability. And we're unique among large platforms and our focus on the safety of users who are under 13, especially given the reality that a large number of young people under the age of 13 have access to phones and to these other large platforms. And these platforms typically aren't designing safety systems for those under 13.

However, as a result of this, we do expect to see continued short-term bookings headwinds as a result and this will lead to a revision in our full year guidance, Naveen will discuss more in his remarks. Now to address the short-term friction, we are taking a number of actions across several key areas. Age checking is our vision of the future, and we believe this tech will continue to scale across the industry. We expect to see continued adoption by other companies in the gaming, social networking social media and AI chat bot spaces.

Through the end of Q1, 51% of global Roblox DAUs have aged checked. As we work to set the global standard in safety full adoption of age check will take time. We expect our new age-based account framework to drive an increase in age check penetration rates and we're focusing on additional means to drive our percent of users that have aged checked to a level that long term, we hope to bring above 90%, which then also unlocks our ability to significantly improve our native communication features. Communication engagement is fundamental to our user engagement and retention and following global rollout of a check to access chat we do see new opportunities to enhance communication features and boost engagement through our higher confidence user age data.

Over the next few months, we're going to be rolling out several enhancements designed to increase communication adoption and improve the chat experience on Roblox. This includes global chat, which is going to allow players across multiple servers in the same game to communicate in a single shared room, which we believe will increase in-game chat density and we're also going to be integrating party chat directly into the in-game chat window, which we believe is going to remove the need for users to leave an experience to coordinate with friends. We plan to expand party chat between trusted friends when appropriate age and necessary parental consent is granted to include voice and Avatar video. And as of now, we're not sharing a ship date for this. We've also planned a system for preset messages that will enable all users to easily coordinate gameplay. Collectively, these investments paired with our incentives for age check will be engineered to drive on-platform communication beyond our precheck levels and deepen user connection. We continue also to evolve our discovery system. Last year, we enhanced really the discovery of long-tail content, and we improved content diversity we're now focusing on high-quality games with beer long-term engagement, and we're currently experimenting with enhanced discovery algorithms designed to optimize for 28-day retention and beyond. Some of our creators have already taken notice as I tweeted this week, and we're implementing changes on how to really create a market that levels the playing field for high-quality, long-term experiences. We continue to see an acceleration of AI tool chain use by our top creators. Nearly half of the top 1,000 creators on Roblox now leverage either Roblox' assistant which is our own AI entry point or MCP model context protocol to compress dev time lines. In addition to our own assistant AI, our creators are using quad code, cursor, codecs and other third-party tools tightly with Robo Studio. These technologies are being used to support creators on everything ranging from light assistance to fully buy coated content. We ultimately leave gaming is a much more complex base than the coating space and really we're driving to achieve iterative Wiggans loop style programming because games are not just based on code but 3D assets, NPCI, core gaming loops and live ops, among others, -- we really want to get to the point where Dev can have Robot studio working overnight on their game and come back in the morning and see those improvements. We really want to try to do a game creation, what tools like codecs, Cursor and Quad code are doing for coding, which is really radically accelerating speed -- and we believe we're uniquely poised to innovate and solve this given the integrated architecture we have from our cloud to our engine to our developer tools to our native AI -- in April, we call that the month that Roblox Studio went Agentic. -- creators can now engage in focused conversations with our assistant about the design and implementation and test plan of new features. Assistant executes the plan now, launches a suite of building agents to test and really try to deliver new features with minimal creator engagement.

We're also on the 3D side to complement coding introducing new mesh and procedural model generation capabilities, which are going to allow creators to build richer world. And we've introduced NPC, which is noncleared character testing agents that can now navigate complex 3D worlds and execute gameplay actions as part of really game development. The vision for MPCs really at Roblox goes beyond PCs, you might interact within game or MPCs used for safety. And part of that vision is really for testing. We believe MPC testing agents can be part of a foundational model that we're building to really help creators iterate quickly on their games using MPCs in addition to humans for testing. And all of this really is to enable small teams to produce super high-quality content very quickly.

We do believe AI is going to fundamentally accelerate gaming, and we are in a unique position to lead in this transformation. Yesterday, we shared on our blog our vision for the future of really integrated AI and gaming and the creation of photorealistic multiplayer gaming and creation that's easy for anyone to participate in. This is really our most ambitious technical innovation to date. We call it the Roblox' reality project. And this patent-pending architecture integrates hyperscale multiplayer stimulation with our current Roblox cloud engine with photorealistic rendering and persistent world state into really a hybrid unified architecture built on our global edge cloud and interim.

The goal of Roblox reality is to really enable creators to construct interactive environments that are high fidelity. And really drive this new technical frontier, the ultimate introduction of the first easy-to-use multiplayer photorealistic platform that we're uniquely poised to build. More details on the blog post, if you want to read up on it. In general, on the AI side, just to wrap up, we have over 400 models running over 1.5 million inferences per second on-prem and on our cloud. This is powering in addition to creation and Roblox reality, everything from discovery recommendations to communication safety, marketplace recommendations and 3D generation and we're now investing in 4 in-house proprietary models, 40 generation NPC behavior, our just shared video super up sampler initiative and coding assistance and generation.

Finally, we've announced several initiatives on our novel games initiatives as we shared, there's a very large opportunity for us in the over 18 cohort. Over 18 represents 26% of the DAUs who have age checked. And once again, in the U.S., DAUs in hours within our 18 through 34 cohort grew over 50% year-on-year. Today, we announced an increase in the Devex rate for age checked 18 users in the U.S. starting on June 8, created earnings for in-game spend generated by age checks, 18 and up users in the U.S. will increase from 37.8% and -- up to 37.8% from 26.6%. I will highlight what made this possible is age check.

Games to do this must meet our definition of novel games which means they're going to utilize our R15 Avatar framework. We're also now working with several well-known game studios to bring reimagined versions of their beloved mobile games to Roblox more details on this in the coming months. We also have announced an internal jump-starting incubator program to support internal existing creators with novel game creation, the initial response to this has been amazing, was strong. And with our existing creators right now with our white glove service, we have roughly 100 novel games being onboarded into the program.

I recently created in Q1 that we've delivered a large number of technical milestones that are all designed at enhancing realism and expanding creator capabilities in parallel with the longer-term Roblox reality project. These are all designed to support high-quality immersive experiences that scale dynamically from 2 gigabyte Android devices all the way up to high quality on a gaming PC. We've introduced our new R15 + Avatar framework with Dynamic heads and this provides a foundation for more lifelike avatars on the platforms, just like recently released Avatar makeup. Together with some of the deep tech like instant streaming, mesh streaming, texture streaming, server authority and slim, these are updates that are removing the barriers that our creators have seen in creating more and more original experiences for older users. We're committed to the long view, our confidence in the future of gaming on our platform has never been higher. And with that, I return and turn the call over to Naveen.

Thanks, Dave, and good afternoon, everyone. I'm going to share a few observations about Q1 and then discuss the changes to our guidance. So with respect to Q1, as Dave highlighted, we saw strong top line growth of 43% in bookings. I think that demonstrates the ability of our platform to grow at a very healthy rate without the benefit of viral hits. We also saw an improvement in content diversity, games outside of the top 10, saw a 43% growth in engagement, 41% growth in spending and as a group, those outside the top 10 games accounted for 65% of the growth in spending. We think this is a healthier level of concentration than what we've seen in the recent past. .

DAUs did come in weaker than anticipated. We'll talk about that more in a minute. But very importantly, I want to point out that user metrics like engagement and monetization remained stable relative to the year ago period. Now as Dave pointed out, we made a number of important safety-related changes to the platform starting with the age gating of comms in January. You'll recall that in our last call, we noted that we expected some headwind to engagement and bookings as a result of the rollout of age checks. We now better understand what I call the second order impacts of reduced comp engagement on things like word-of-mouth and organic content growth.

Additionally, the monetization bias of our recommendation engine likely negatively impacted App Store ratings and ultimately, sign-ups. We do have additional safety features, including kids and select accounts rolling out later this year. Those have huge long-term benefits. We've always viewed safety as a compounding moat for Roblox, and these features are an important ingredient to that. But it does mean continued friction in the short term until we get the benefit of continued adoption of age checks, the planned updates to our comms features that Dave described which we believe will improve chat vitality and chat density and discovery enhancements that result in better content recommendations.

As a result, we are lowering our guidance for full year top line growth to account for a continuation of the safety headwinds that we've experienced to date. Our revenue guidance for the full year will now be 20% to 25% and our full year guidance for bookings growth is 8% to 12%. That guidance is based on the expectation that DAUs will continue to contract between Q1 and Q2 and then return to sequential growth in Q3. And consistent with our prior guidance, we do not assume any major viral hits in those numbers. The reduction in our bookings expectation will also impact margins this year. As you would expect, much of that is related to fixed cost deleveraging given the change in our bookings expectation.

Although roughly 1/4 of the margin reduction relative to our prior guidance is related to the incremental investments we're making in AI and the 18 Inova DevEx increase. Now even though those investments are incremental to our prior margin expectations for this year, in future periods, we plan on them being funded by the capture of additional operating leverage. And even more importantly, we are highly enthusiastic about what they can unlock in terms of long-term growth, which, of course, continues to be our North Star. With that, we'll open the line for questions.

[Operator Instructions] Your first question comes from the line of Eric Sheridan with Goldman Sachs.

Maybe a 2-parter, if I can. With respect to incenting the development out of the 18 plus community, I wonder if you'd go a little bit deeper into signal you were getting in terms of that type of content from developers in terms of what it might mean for the long-term health and compounded growth longer term for the business? And second, with respect to the change on DevEx, maybe just help us better understand why that was the right number to move to in terms of higher DevEx

Great question. There's a big invention on Roblox a while back when we moved from a platform without Roblox to DevEx, and we saw immediately that the incentive of creating a closed-loop ecosystem where robots could be used by our users devs could build interesting experiences. They could then cash those out, created a virtuous cycle, really, that's been somewhat of a machine-driving Roblox ever since. And we've seen massive organic behavior from that virtual economy -- we've been very careful in increasing the Creator DevEx rate over time, but we have slowly increased it.

As we did at RDC recently, the 18 and up market now that we have age check, we can know quite accurately who's a real 18 and up player. And in addition, that market globally is roughly 80% of the global gaming market, which is somewhat astounding given our low actual penetration, although quickly growing penetration. Those users also, as we shared, monetize at 1.5x where we are today. And that's a big part of our future in addition to our existing [ U1 ] base. So we know systems drive behavior. We want to see novel games on our platform. We want to make it very profitable for creators to make amazing content -- we want them to trust our discovery systems, we'll organically reward experiences with long-term retention. And so this is just a continuation of Roblox being a systems company. focusing on this big area we're moving towards.

Next question comes from the line of Matthew Cost with Morgan Stanley.

Great. Naveen, maybe if you could just dive one step deeper into the reduction in the guidance. I think you gave a lot of really helpful detail about the behavior of people related to age check. Is the entire guide down driven by the change in behavior versus Feber that you've observed because of age check? Or are there other material factors that are worth calling out -- that's question one. And then the second is just on Roblox reality. Are there any increased cloud expenses or CapEx that you're expecting to support that going forward.

Yes. Thanks, Matt. So with respect to the guidance, the way that I would characterize that -- excuse me, the change in guidance, the way that I would characterize that -- it is largely safety-related. There are a number of aspects to that, as we spoke about. It's age checking. It has a follow-on impact to comms and that affects both users that have age check as well as those that have not. And there's a number of things, as Dave laid out that we're doing from a product perspective to reignite comm's engagement on the platform. We do also believe that some of the dynamics that we saw around discovery being sort of more monetization biased than we would like plays into that. Content recommendations were probably not as optimal as we would like to see them. So I think it's really the combination of those dynamics, but largely driven by what we're doing from a safety perspective. In terms of Roblox reality, Dave, did you want to comment on that?

First off, robots reality will not be free. The type of technology we're showing here is going to combine Roblox hybrid engine in cloud, which I think if we you look into all of our financial statements, you can see we run at less than [indiscernible] per hour roughly. But simultaneously, what we see as the future, we're building close to photoreal or photoreal experiences, especially when multiplayer is very difficult. It takes enormous production budgets and part of what Roblox is about is democratizing creation for everyone. We do believe this is the ultimate combination both traditional 3D network, multiplayer gaming engines like Roblox, which is somewhat unique, given its cloud with all of the future research we see in video models, moving more and more towards real-time. We have the opportunity to build a very Roblox specific video world model that we call a super sampler that can key both on video as well as 3D spatial information to do arguably what we believe is going to be a beautiful job of up sampling under developer control with developer prompts.

But I do want to highlight we're right on the edge really in the whole AI space of running real-time photoreal video models 2K at 60 hertz. And that's why we say this is an up-and-coming project. When we launch this, this will not be free. This will use cloud compute, we will have some kind of way of subscribing or paying for this. And because of that, I think we will offset the real-time inference side of it. On the training side, we may have additional things. I don't think we are fully showing them this year, and Naveen may have a compliment on that.

Yes. Just to put a finer point on the expense side of it. We've not changed our expectations for CapEx this year. We think that a lot of what we're doing in terms of landing GPUs in our data centers will cover what we intend to -- or what we're going to need this year. There is some cloud training that we will be doing. I mentioned that, that is now factored into the updated margin guidance. And as this technology continues to develop, we will use a combination of cloud as well as our own data center capacity to execute both training. And as Dave said, inference will be something that will be funded by usage.

Your next question comes from the line of Ken Goralski with Wells Park.

Two, if I may. Maybe first -- could you maybe Naveen, just talk a little bit about what you've seen -- I mean, I'm a little puzzled on 90 days ago, big guidance of '22 to '26 and that you gave the first quarter guide, which you came near the high end and now we're -- the guide is meaningfully cut for the full year. Again, I'm focused on the bookings side. Could you just tell us -- I mean, how much of this is impact that you're seeing kind of already in the 2Q period versus the expectation of some of the changes that are rolling out in the June period.

And then maybe secondly, for Dave, please. Could you talk about like just philosophically, the changes you're making in terms of what content can be discovered by younger users. There are some changes that involve kind of embargoing of sudden content. How does it fit with your notion of democratizing kind of publishing content and discovery. There's kind of -- it feels like a little tension between content safety and moderation and kind of democratizing the platform. Could you just talk about where you could see the balance of those 2 things today.

Ken, I can start on the first one and then hand it over to Dave. So look, we launched age check in January. And as we said, we expected to see some headwind both in terms of hours and DAUs. I think what we did not fully understand until we have now the benefit of 3, 4 months of experience is how that impacts the platform more generally with respect to comms engagement and the knock-on impacts from that. Now I want to point out a couple of really important things: Number one, when we look at what has happened over the last few months, as I noted, engagement has actually remained quite strong. Monetization has remained strong and retention has remained strong. What we have seen is challenges at the top of the funnel, i.e., new users coming in.

And the reality is when we think about the rest of the year, we're not going to see the bookings impact of that right away, hence, the performance in Q1. But we do know that the fact that we had more sign-up headwind over the last few months is going to put pressure on bookings over the remainder of the year. However, when we start to get back to DAU -- sequential DAU growth in Q3, given the strength of monetization and engagement we feel confident that we will be able to drive the bookings growth that we're guiding to, which given the comps in the back half of the year, equates to something like relatively low single digits. So we're not trying to do something heroic in the back half of the year, and it is largely driven by return to DAU growth.

Great question on democratizing really creation. Two metrics you can look at Roblox at over many of the past years. The first is total dollar value creators participate in, what's the size of Devex every year, that's growing very rapidly. And then the second is the growth rates of Creator [indiscernible] and consistently the number goes up the total amount and the growth rate of the Creator 1,000 is consistently faster than 110 or 1, so that's a metric to measure democratization. We believe -- I believe that will continue. Relative to what we've done with kids accounts and select accounts, we believe we're striking a very good balance, and we have this ability to strike this balance because we age check.

Other platforms that don't age check have a wide range of users signing up at various ages. We know what age everyone is, and it more allows us to align the appropriate content with them. We are doing that with kids and select accounts. The corpus of that content is going to be very, very, very large. It's going to cover the significant proportion of what they play today and it's going to allow creators to participate in a UGC space for 16 and up to introduce innovative concepts and to allow us to be very careful with what content reaches under 16. So I'm optimistic you're going to continue to see that democratization.

Your next question comes from the line of Cory Carpenter with JPMorgan.

Thanks for the question. I had 2 that are related. Naveen, you mentioned a few times now, expectations to return to DAU growth in 3Q. I understand it's a seasonally strong quarter, but you're also rolling age-based accounts out in June. So maybe just talk a bit about what's giving you the confidence in the return to the growth that quarter. And probably on a related point, could you just elaborate a bit on the changes you're making to address some of the communications friction points you've seen and what the time line of rolling those out is.

Yes. Thanks, Cory. I'll hit the first part, and then I'll let Dave talk about some of the changes we're making on the product side. So look, the reason that we are speaking to DAU growth in Q3 are a few things. Number one, as you pointed out, it is a seasonally strong quarter. So we expect to continue to get some tailwind from that. Number two, we will have a number of the product changes that Dave is going to touch on in a second rolled out by that point in time. Number three, to your sort of speculation around the impact of kids and select accounts during that period of time. We don't expect that to be as dramatic as what we did in January with respect to comms. And what I mean by that is when we started age gating access to comms, it was a binary experience for our users. They lost complete access to comms. And as we've spoken about, that had a number of knock-on impacts in terms of overall sort of vitality on the platform, sentiment, app store ratings, et cetera.

What we're doing with Kids and select accounts will have some impact. So it's not sort of immaterial. However, it's not as big a change because kids are going to come in or people who have not age check, they will still have access to 20,000 games that represents more than 97% of the engagement on the platform from those cohorts. So it's not nearly as sort of drastic of a change as what we saw with age gating comp. So that's kind of how we're thinking about the impact of all this stuff over the next couple of quarters. Dave, do you want to touch on the comms road map?

Absolutely. So some of these things are in the pipeline and live and experiment. So I believe global chat is actually live in an experimental mode with a subset of our users right now. And I do want to highlight this is a way to give that feeling of chat density even if a subset of people are not aged check and not on comms because this will allow people playing the same experience to connect with people close buying a separate server. Also, with preset messages, this is a way to support what is very common in Roblox games, which is gameplay coordination, not necessarily full on chat that we believe can safely be used across age bands. This is in the hopper pretty soon. And then the bigger initiative is really to move party chat fully natively within experience, arguably played on second device on a lot of the common functionality that friends use when playing Roblox either tech channel or voice channel side-by-side. We're not giving a date on this right now. But this is, I would say, the third top priority for our comms road map.

Your next question comes from the line of Jason Bazinet with Citi.

I just had a quick question on the over 18 sort of DEVEX incentives you're giving, what have you learned in terms of when you roll out incentives like this in the developer community response and then the users respond to the developer community, how long does that gestation period take? I know the question but sort of spring yes.

It's a great question. This is a board-level discussion prior to doing Devex at all where there was a time where Roblox was a hobby for many. And as part of that hobby, there was no way to make a living on Roblox much less creative studio that makes $10 million or $20 million or $50 million. What we have seen over time is the second we introduced Devex and Roblox the quality of the experiences on Roblox got much greater, much more quickly because creators were able to make that a full-time job, dedicate their life to that, and that's why we've evolved to the point where we literally have thousands and thousands of people who make their living on Roblox.

The other thing we've seen over time is we have continuously made incremental updates to the DevEx rate. It started very low. It's climbed. We introduced it -- and so we have a fair amount of anecdotal evidence that when creators can trust, they will make a certain amount of revenue from their experience they will commensurately put more effort into the creation of quality experiences. So it's really been part of the Roblox history. We've seen it time and time again pay off in higher quality content that creators can trust to build.

Your next question comes from the line of Andrew Marok with Raymond James.

I was wondering if you could give a little bit more detail on the shift in discovery algorithms that you had mentioned in the prepared remarks, going from some of the shorter-term monetization factors to the longer-term health factors. How does that maybe evidence from a user perspective? And then how does that maybe encourage that increased chat density over time?

Yes. I would say from a user perspective, this is something, I believe, our creators and ourselves and me included, can intuitively literally see on our home page which is the quality of -- and the types of experiences and the ratings they may show. And our creators have noticed this. We have -- we've made great leaps in discovery in using ML to predict many factors for each user game pair in the whole system. Now that we've developed this capability and this muscle, if you will, we have analyzed where we are and the potential for 18 and up and made the determination we really want to drive the user growth of that 18 and up segment even if we are slightly less weighting short-term monetization.

So we can see it intuitively, and I would say our creators have noticed as well and our creators welcome it because they want to really invest in long-term retaining high-quality properties and somewhat frustrating for them when they might see a shorter-term monetization type game that really isn't going to be around for a long time. So I would say both intuitively as well as systemically, we believe it's the right thing to do.

Your next question comes from the line of Omar Gusek with Bank of America. .

So I wanted to point out that, I think, in the third quarter of '25, your European and U.S. daily active users were about 60 million, in the first quarter, it's 51 million. I think you've characterized in the past that the kind of the viral surge in the third quarter was a big user acquisition event and that a lot of your users look the same. So I wanted to know if you would still kind of consider that as a high watermark that you can get back to if you simply get through these growing pains that you talked about today or if, in fact, that the retention that you thought you had will not materialize? .

Yes. Omar, it's Naveen. I'll take that. First of all, just to put into context some of what you're seeing with respect to the Dow trend for Europe and U.S., don't forget that there's roughly, I think, 4 million dows that came out due to the Russia block. So that's an important thing just to note. That being said, sort of -- with respect to, call it, the broader question, we expect that we will return to DAU growth as we get through some of the safety friction on the platform. What we've said in the past is that the users that we acquired through viral games last year had similar retention characteristics to, call it, the platform at large. That continues to be true. As I mentioned a few minutes ago, when we've looked at what has happened over the last few months, sort of the nature of this friction, retention has remained strong. The reason that DAUs have come in weaker than we expected, is largely top of funnel sign-ups, which we believe is related to what we've seen in terms of comms friction and how that's been reflected in organic growth through the app stores.

Your next question comes from the line of Ryan Pete with BMO Capital Markets.

This is David Lusoga for Brian. Dave, you talked about working with some well-known studios to bring their mobile games into the Roblox ecosystem. -- on the conversation of higher Devex for certain things like over 18 cohorts, can you talk about your expectations for how the revenue share with some of those creators might compare to the broader mix shift revenue to the platform?

Hey, we do everything linear and we do everything as a system. So I believe it's safe to say I'm looking around the world, we've never cut a custom revenue share deal like -- we've had lots of discussions in the past about accelerators and de accelerators and all of that. But Devex is consistent across everyone. The benefit of raising it for 18 plus is those studios we're working with are going to get that as part of it. And one of the most attractive things to the studios we're working with, whether they are traditionally a mobile game creator, or a PC game creator, is there's something very unique about the Roblox architecture in that it is the same build, the same game, whether it's PC, tablet, console or PC it's dynamically scaling with a lot of the tech we mentioned, whether it's slim, streaming, textures, streaming, mesh and so there is the ability to unify a single build for both of those platforms, which is very attractive to a lot of the studios we talked to.

SP255882563 Okay. So I'm getting the queue. So we're going to wrap up here. I appreciate all of your questions and feedback, and I'll kick it back to Naveen to see Naveen, if you have any closing remarks.

I just want to thank everyone for joining us this afternoon, and we look forward to speaking to you next quarter. .

This concludes today's conference call. You may now disconnect.

All right. Good morning, everyone. Welcome to day 3 of the Morgan Stanley TMT Conference. My name is Matt Cost from the Morgan Stanley U.S. Internet team. Very excited this morning to be joined by Naveen Chopra, CFO of Roblox. Thank you for being here.

Thanks for having us.

Yes. For research disclosures, please note that all important disclosures, including personal holdings disclosures and MS disclosures appear on the Morgan Stanley public website at ms.com/researchdisclosures or at the registration desk.

All right. And with that, let's start with the hot topic. The market clearly views interactive entertainment as vulnerable to disruption from AI. And the technology is evolving so quickly that I think it's an understandable question investors are asking about the potential for disruption. So let's just start there. What do you think about the way the market has been approaching that topic? And how do you think it applies to Roblox?

Yes. Well, thank you for going straight there. No beating around the bush. Look, I don't think it will surprise anyone to hear if we think the market has it wrong. We think AI is very much a tailwind and an accelerant for our business. It's one of the reasons that we've been investing in it aggressively over the last several years.

And I think the reality is like if I step back, the market seems to be kind of assuming that you're going to have this very small handful of AI winners and everyone else's business is going to get vaporized.

And I think that's honestly an overly simplistic way of thinking about how new technologies tend to be adopted and ultimately commercialized. There will be companies not named OpenAI and Meta and Google who use specialized AI to disrupt their competitors and ultimately become winners in the next generation of AI and technology.

And that's very much how we see Roblox, and we are mobilizing the business aggressively to make sure that we are one of those disruptors in our space. We have a large-scale platform business with very strong network effects. And we are using and building gaming-specific AI to further entrench our position as the platform of gaming.

We see a world where these AI technologies will allow people to someday create games without having to write a single line of code. That translates to more creators and more content. And if you run a platform business like ours, that makes everything that we do from the discovery platform, the monetization platform, the safety infrastructure, the hosting infrastructure, the communication capabilities, the social graph, it makes all of those things much more valuable because you got more content and more creators. So we're really excited about what this is going to unlock for the business. And as I said, we're investing to try to make it happen.

Great. So what does it mean to be a disruptor, I guess, following on from that? Because you have a lot of tools that you've announced already like the Cube model and the world model and other content creation capabilities in Roblox Studio. So what are they enabling that wasn't possible before? And what feedback are you getting from creators and users on those?

This space is obviously moving very, very rapidly. And one of the ways that we have started to think about our product road map, call it, for the short to medium term is to do for gaming what we see a lot of other software development tools doing in, let's say, software more generally.

Let me explain what I mean by that. I think most people here are probably now aware that over the last, call it, 3 to 6 months, we have seen some -- this massive acceleration in agentic coding. And what these agents are able to do in terms of software development is like mind-blowing.

These things code 24/7. They are infinitely scalable. They're iterative, so they code all night. They test their own code in the morning and then they fix a bunch of stuff and make it even better the next day. And they are able to do that because they are built on top of very sophisticated, very capable operating systems with APIs and the like.

In gaming, Roblox is that operating system. And so we are starting to now provide the ability to plug in both AI tools that we develop internally as well as third-party AI tools to create that sort of acceleration on top of our own platform. And we're actually starting to see developers experiment with this kind of stuff. And they are using these tools to design games, to build 3-dimensional assets to orchestrate gameplay. And it's pretty exciting when you see what they can do with it. In fact, there were some folks just yesterday who were showing me a demo. These are folks internal at Roblox, where they had taken some of these AI agentic tools, again, a combination of stuff we've developed internally as well as some external stuff.

And they created a game where you could drive around the streets or actually walk around the streets of San Francisco. And the whole thing was prompt based. So they basically said, create a game where I can walk around the streets of San Francisco, make it physically accurate and as high fidelity as possible.

And in 24 hours, they had a game created entirely by agents where you can walk around the city of San Francisco, you can look at -- it's basically the downtown area, but you can look at the buildings. They're physically accurate. Every traffic light is in the right place, 24 hours.

And that ultimately is going to turn into a world where you could have a team of, call it, 3 people developing like a GTA class game in maybe a week. Not there today, but that sort of acceleration in content creation is what we are so excited about and what we think will unlock tremendous growth in our business.

You mentioned in there, this kind of a mix of internal and external tools. I guess what is the mix between those 2 things? And how do you approach the decision about whether to build from new, whether to use open source or whether to work with a larger partner for those tools?

The reality is developers are going to make that decision. What we -- our goal is to make sure that we have all the tools available that allow you to make really high-quality games. At the end of the day, we're ambivalent as to whether you use our tools or whether you use third-party tools. We want that content to exist because that's going to make our platform more valuable.

That's like the MCP server, I guess.

Exactly.

Allowing that interoperability. Got it. So I think moving away from AI for a moment. The platform has seen a lot of growth in users and engagement over the past couple of quarters. There are a number of really notable hit games on the platform last year. So on that topic, I want to start by asking what has gone right? And how do you think about the sustainability of that engagement as we move through 2026?

Yes. I mean, as you pointed out, '25 was an incredible year for us. Just to give people some dimensions around that. I mean if you look at users on our platform between the end of '24 and the end of '25, we grew users by 70-ish percent and bookings growth was 55% for the full year. So very healthy growth. Now there is no single sort of silver bullet that I would point to and say this is the thing that we did that unlocked all of that.

The reality is this is a platform business, and there are initiatives across every dimension of that platform that work together to enable that kind of growth and enable the success of some of the very big viral hits that obviously contributed to 2025.

Just as a few examples, we continue to make a lot of progress around things like search and discovery. That led to significant sort of content diversification on the platform. We saw a double-digit increase in the number of unique experiences that people engage with in any given month.

We've also seen double-digit increases in time spent on a lot of these different experiences. So real benefits coming out of continued evolution of search and discovery. You look at monetization, and we have continued to deploy things like regional pricing and price optimization. These are basically tools that help our developers ensure that the virtual goods that they are selling on Roblox are optimally priced given the nature of the good and the location of the user.

And as we've continued to evolve those kinds of capabilities, we've seen, for instance, outside the United States, massive growth in the number of payers on our platform, basically doubled the number of international players on the platform. And we've also seen very, very healthy growth in bookings per active user.

So all those things work together. We've made major strides around safety as well, which is a huge priority for us, as you know, most recently rolling out facial age estimation. And so it's a combination of all of those things that really facilitated the growth that we saw in 2025.

And we expect to see more of that in 2026. We've got a lot of momentum coming out of 2025. We like what we're seeing in terms of the content pipeline. We've got a really ambitious product road map. And it's the combination of all those things that make us bullish about the long-term growth opportunity in our business.

There are a lot of things in there, and I noticed one thing that you didn't raise is, oh, we're going to have another giant viral hit in 2026. Obviously, those things are hard to predict, but we saw it happen a couple of times in 2025. I think there's a perception from investors that Roblox is a little bit of a hit-driven business. How do you think about it? And how would you respond to that?

Yes. Well, I'd respond, I guess, by pointing out some important facts, right? I mean, Roblox is a platform with over 14 million different experiences, millions of different creators, 144 million daily active users as of the end of '25 and call it, 125 billion hours of engagement.

And interestingly, if you look at 2025, on average, I think something like 2/3 of that engagement time was spent in experiences outside of our top 10. So what we describe as kind of the torso and the tail of the platform is a huge part of this business. And it's growing at a very healthy rate. I sort of highlight these stats on each of our earnings calls. And like in Q4, those experiences outside of the top 10 grew bookings at, I think, like 55%, 56% year-on-year.

So they are really important parts of the platform. And if you look at the business as a whole, like in Q4, we didn't really even have a big new viral hit. And you still saw very high levels of growth in users, in hours, bookings grew 63%. So I obviously came from the entertainment industry. I know what a hit-driven business looks like. It's not what I see in Roblox.

We love the big viral hits. They obviously contribute to the platform. They help attract a lot of new users. They lift other elements of the platform, but this is not a business that's dependent on those.

I think a closely related concept to that is genre expansion and trying to broaden what's available for people to play on Roblox. Maybe give us an update on the initiative that you've been undertaking for a while now to broaden out into many different types of games and experiences and sort of what opportunities it opens up as you try and continue to reach a wider and wider audience.

Yes. Look, for us, genre expansion is really a critical part of audience expansion, right? We want to continue to attract new users to Roblox. And one of the things that's been really exciting for us that we've learned over the last couple of months really just as we've rolled out mandatory age checking, which now gives us our first view of like the actual age of people on our platform.

And it demonstrated that we have a huge opportunity ahead of us to bring older users on to Roblox. I think we're like less than 10% penetrated in 18- to 34-year olds in the U.S. market. So there's just massive opportunity ahead. And we know that we need to evolve the selection of content that exists on the platform in order to attract those users. And so -- there are a variety of things that we're doing to try to make that happen.

I would largely put it into kind of 2 different buckets. First bucket would be a bunch of stuff that we need to do from a tech perspective, right, to be able to facilitate games that have higher fidelity, higher performance, more complexity. We've talked and demonstrated some of those things in the past, whether it's Texture Streaming or SLIM or Server Authority.

We won't get into the weeds of all of those things, and there's a whole road map of stuff that we'll continue to do to facilitate those kinds of games. And that all gets accelerated by the AI stuff that we were talking about at the beginning. So that's really exciting.

And then from, call it, a platform perspective, there are a variety of things that we are starting to experiment with to really help our creators be successful with different types of content on the platform. And that's anything from thinking about our discovery experience, our user interface and perhaps targeting that or designing it a little differently depending if you're a 14-year-old user or a 34-year-old user.

Thinking about the promotion algorithms, the off-platform user acquisition that we partner with developers on, the financial incentives that we provide developers in terms of their economic participation, like all of these things are levers that we can pull to try to encourage development of content that we think will appeal to this older audience, which, as I said, is a really exciting growth opportunity.

Got it. Maybe translating some of those themes into your outlook for 2026. I think that where -- your guidance for 2026 bookings, I think, was comfortably ahead of where the buy side was expecting.

You saw the stock react positively to that. Maybe tell us more about the work you did internally to arrive at that, the underpinnings of, I think, 22% to 26% bookings growth this year. And since you won't be providing annual guidance after 2026, I guess, how is your guidance philosophy evolving now?

I'd say there are probably 4 things that we were focused on in thinking about how 2026 would play out. First, users. We had massive user growth in 2025, as we've talked about. And we wanted to make sure we understood what the behavior of those users was going to look like because they represent a huge chunk of the platform.

And we've been encouraged by what we've seen. Retention into, call it, month 3, 4, 5, 6, which is kind of really important in the business has been healthy. The engagement and monetization of those users looks a lot like users on the rest of the platform. So we can kind of think about this as a lot of the historical trends continuing. So that was an important input. Second, content. As we talked about, the viral hits are really difficult to predict, and we don't bank on those in terms of how we put together our guidance. But we do assume that there's going to be a cadence of new fresh content on the platform. And so we look at that pipeline, and we've been, again, encouraged by what we've seen in the early part of 2026.

And that is important to really that core of the platform, right, which is the 2/3 plus that is not in those top 10 hits. So that's something we've looked at closely. We've also been very focused on the rollout of facial age estimation. In early January, we basically started requiring people to go through an age check in order to access communication features on our platform. It's very important from a safety perspective.

We knew that, that would create some headwind in terms of engagement and bookings when we first rolled it out. It would take time for people to go through that process and then kind of return back to earlier levels of engagement. And so we took some time to study -- we rolled that out a little earlier in a couple of territories, and then we rolled it out globally in early January. And so as we got more data about exactly how that was impacting things, we were able to factor that into our guidance as well.

And then the fourth element is our product road map, and I would not underestimate the importance of that. We have a whole bunch of initiatives across the engine, across user, across communications, across discovery and very importantly in our economy as well.

And we actually work very hard to estimate how those things are going to impact bookings growth and engagement growth on a year-over-year basis, and all of that stuff gets rolled into our guidance for the year.

Got it. You talked about facial age estimation is kind of one of the important assumptions underpinning it. And you've mentioned even in this conversation, how important safety is. It comes up a lot around Roblox. I guess, tell us more about why that's so important. You mentioned how you've kind of baked in the headwind to engagement into your expectations. But I wonder as you look out to the end of the year and into '27 and beyond, are there positive opportunities that could come from knowing how old your users are?

Absolutely. I mean this notion of facial age estimation for us started with safety because we view safety of kind of everything on Roblox. Like if you want to provide a safe experience for people of all ages, you actually need to know the age of the people that are on your platform.

And unfortunately, the reality is until very recently, that was a difficult thing to do. And it's the reason why the vast majority of large-scale digital platforms rely on self-reported age information. And guess what, when you ask people to self-report their age, they lie. It turns out they lie in both directions, by the way.

Young people lie about being older and older people lie about being younger. I guess no one is happy with their current age. But knowing the age of our users allows us to both provide a safer environment, but also just a better experience for our users. So from a safety perspective, it's about doing things that make sense from, call it, a societal norm perspective, right?

So if we know you got a 12-year-old on the platform, we're generally okay with that person being able to communicate with an 11-year-old and a 13-year-old, maybe even a 14-year-old. But it doesn't feel right for that person to be communicating with a 35-year-old that they don't know in real life.

So we need to actually know the ages of these people in order to put the appropriate guardrails in place, and that's what we're doing now with the rollout of this technology. But it also makes the product better. And what I mean by that is, again, if you just stick with the communication example, if I'm going to put you in a game, chances are if you're a 12-year-old player on Roblox, that experience is going to be a lot more fun if I put you in a game with other people who are also 12 years old. The way they play, the way they communicate, the lingo that they use it's on chat, guess what, it's different if you're 8 years old versus if you're 13 years old versus if you're 25 years old. And it's just a better, more fun experience for the players if they are engaging with people sort of in a similar age band.

And that extends beyond just the communication piece of it, the content that I surface, the way that I think about what I'm going to promote in the user interface. If I know you're 12 years old versus if I know you're 25, I'm going to do different things. So having that information is incredibly valuable for us, both from a safety perspective and also just from making a better product experience.

Got it. So it sounds like my days of going and beating middle schoolers or front lines are a little numbered then.

Hopefully.

So direct payments. So it's something you called out on the fourth quarter as scaling in the mix of how consumers are buying Roblox. So I'm wondering what are you seeing in terms of the financial impact there? How do you think about the longer-term implications for margins as direct payments become a larger part of the mix?

Yes. Look, we've seen meaningful operating leverage capture by driving down COGS rate, moving more of our payments to web platforms and other lower-cost platforms. Gift cards are a big part of our business as well, and they tend to have lower cost than app stores. And that contributed meaningfully in 2025. I do expect it will continue to be a contributor in 2026. And one of the things that I pointed out on the call is that we've got this headwind to margin in '26 from the DevEx increase that we announced in 2025. But if you look at like at the high end of our guidance, we expect margins to be roughly flat despite that headwind. And one of the reasons we think we can get there is because of incremental leverage that we expect to see from COGS.

That's not going to be sort of linear improvement every single year because there are a lot of different things that go into it in terms of how aggressively we push people to different platforms, what we can negotiate with different platforms. In the gift card part of the business, there are different fees between physical gift cards and digital gift cards.

But we see all these things moving in the right direction such that COGS will continue to be a source of leverage for us and one of the things that we can use to fund either margin expansion or increases in DevEx for our creators.

And how do you approach the decision between those 2 choices?

Look, in the long run, they're both critically important to us. We want our developers to have an even more exciting economic opportunity on Roblox. And we also are committed to expanding margins in the business over time. So the trick is that there's a little bit of a sequencing dynamic between the 2, like we like to capture the operating leverage and then use it to fund other parts of the business, if you will. So generally, that's how we try to approach it. But we'll see how that plays out over the next few years.

Got it. Let's move on to advertising. So we've written a lot in the past few years about how large we think the advertising opportunity on Roblox might be, especially as you look out over the next couple of years. And you've had some important developments there just in the past year or so in terms of rewarded video going GA.

You continue to make progress on Sponsored Tiles. I guess how is that rollout tracking compared to your expectations? And what do you want investors to know about Roblox as that opportunity?

I think what I want people to know is that we are very excited about the long-term advertising opportunity on the platform. In the short term, there is still work to be done to scale it. I think the long-term opportunity is exciting because we have some really unique assets to bring to advertisers, right?

We have the ability to safely advertise to younger audiences, which I think is becoming more important by the day to advertisers and users. We have tremendous engagement on our platform, call it, roughly 3 hours a day. So this is an audience that advertisers really want to get in front of.

Now at the same time, inserting advertising into gaming is something you have to be very thoughtful about it. It's different than, call it, linear video advertising where you can say, "Oh, every 12 minutes, I'm going to show you an advertisement". With gaming, it's much more interactive. And so you have to sort of be -- you have to work very closely with game developers to make sure that you're inserting advertising in a way that is not going to ultimately be detrimental to the user experience.

So we really think about our advertising business in 2 buckets, sort of a traditional brand advertising business, and then there is an endemic advertising business, which is really what allows our creators to pay for placement and user acquisition on our platform.

We've got really strong product market fit in both of those based on, I think, sort of our early learnings. On the branded side, as you mentioned, we are really focused on our rewarded video product, which has the benefit of being an industry standard product. It's well integrated. It's highly measurable. But it is something that does take some time, as I said, to integrate from a supply perspective.

On the endemic side or for creators, our core product is Sponsored Tiles, which again allows them to pay for, call it, premium placement in our user experience. That business has seen some very healthy growth, again, starting at small scale, but the growth rates are pretty compelling. And we've been able to give our creators the tools to manage those kinds of campaigns and very importantly, to help them measure the impact of those kinds of campaigns.

From a financial perspective, those dollars actually don't show up as bookings. They are an offset to DevEx in our P&L. But it's still an important contributor to the overall economic equation over time. Again, in the near term, contribution is modest, growth rates are high, longer term, big opportunity.

Got it. When we think about the user growth opportunity, domestic and when I say domestic, I'll include kind of North America and that versus international. Where are you in the journey of reaching those audiences in kind of your core most mature markets versus elsewhere? Because you called out on the most recent call, strengthen in Japan, India, Indonesia as real drivers of growth in APAC. How do you think about the opportunity in those markets versus where you are in the U.S. and Canada now?

Well, what I think is super exciting about those markets is just that we're we have very low penetration. So we've seen, as you pointed out, like tremendous growth in the last few quarters in those markets, both in terms of users and monetization. I mentioned earlier that like the number of international payers that we have on the platform kind of doubled in 2025.

So the slope of the curves there is really exciting, but the penetration is still small. So there's a huge growth opportunity. There's also a growth opportunity from a monetization perspective. When you look at how many of those regions monetize relative to what we see in the United States, there's a big gap. I don't expect those things to necessarily equalize. Just there are economic differences between countries. But there's definitely room for growth and a lot of the things that we're doing from a product perspective and a content perspective are intended to facilitate that.

The formula for us to continuing to drive that growth and maximize that penetration is actually the exact same thing that we're doing in the United States, meaning it's about product, it's about content, it's about audience expansion. And if we do those things, it will drive growth in the United States, and it will also drive growth globally.

The beauty of gaming is that it is a pretty global market. So the things that appeal to users in the United States and Canada tend to also appeal to users in other parts of the world. So there's a lot of leverage from that perspective in the business.

Got it. Maybe close out with kind of 2 more thematic ones or longer-term ones. I guess you've talked a lot over the past couple of years about a goal of reaching 10% penetration of the game software market, so around $200 billion. I think you're about 3% to 4% penetrated in that today. So of everything we've talked about today, what are the key levers investors should be watching to measure your progress towards that goal other than, of course, bookings?

Right. Well, by the way, we have an even more ambitious internal target for U.S. market share that we would like to achieve. And we've talked about a number of the different initiatives that we have to try to accomplish that. But if I really boil it down, -- the path to achieving those goals is about tech.

It's about genre/content expansion, which leads to audience expansion, which translates to growth and monetization. Like that's the formula. If we can put all those different pieces together, we will continue to end up this marketplace.

Got it. And then finally, just on AI kind of ending where we started. What would you highlight as maybe when you talk to investors, what comes out as the most underappreciated opportunity to deploy AI at Roblox and take advantage of it? And then what is the challenge that you're most focused on executing through from AI?

Look, the AI opportunity is largely what I said at the beginning. I mean we see this as a massive accelerant to our platform. And I think if you -- the folks who sort of look at it as a threat, don't understand the nature of our business.

As a platform, we want more content. We want more creators. We want it to happen faster, and we can use AI to help make that happen and provide people the tools to make every aspect of our platform that much stronger. So we could not be more excited about it. And if you look at how we are mobilizing internally to make that happen, it's like at the core of everything that we do.

Challenge, look, I think the biggest challenge that we deal with is just sort of the short term versus long term. Like there's just the amount of time we spend people trying to figure out like, oh, what's going to happen to bookings in Q1 or Q2. And the reality is like that has no bearing on what the value of this business is going to be 5 years down the road. So our focus is on the latter.

Got it. Thank you so much for being here.

Thank you. Thanks for having us.

Good afternoon, everyone. This is Tiffany, and I will be your conference coordinator today. Welcome to Roblox's Fourth Quarter and Full Year 2025 Earnings Call.

[Operator Instructions]

Now I will turn the call over to Jaime Morris, Roblox's Head of Investor Relations. Jaime?

Thank you. Good afternoon, everyone. I am very happy to be here for my first earnings call at Roblox. Thank you for joining us to discuss our Q4 and full year 2025 results. With me today is Roblox's Co-Founder and CEO, David Baszucki; and our Chief Financial Officer, Naveen Chopra.

Before we begin, I would like to remind you that our commentary today may include forward-looking statements, which are subject to risks, uncertainties and assumptions that could cause actual results to differ materially from those described in our forward-looking statements. A description of these risks, uncertainties and assumptions are included in our SEC filings, including our most recent reports on Form 10-K and Form 10-Q.

You should not rely on our forward-looking statements as predictions of future events, and we disclaim any obligation to update these statements, except as required by law. During this call, we will also discuss certain non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics can be found in our shareholder letter and supplemental slides, which are available on our Investor Relations website.

With that, I will turn the call over to Dave.

Thank you, and good afternoon, and thank you all for joining us today. One year ago, we shared our goal of growing year-on-year bookings at 19% to 21%, and we talked about our goal of capturing 10% of the $200 billion global gaming content market on Roblox. The results we're sharing today for Q4 and the full year 2025 demonstrate the incredible progress we've made. In 2025, we significantly exceeded our guidance, both on revenue where we grew by 36% and on bookings where we grew by 55% year-on-year. We saw the emergence of viral hits that broke concurrency records at a platform peak in August of 2025, we hit 45 million concurrence on Roblox, and we saw new records for individual games on Roblox, including in September of 2025, Steal a Brainrot peaked at 25.4 million concurrent players at the same time.

We saw strong engagement and bookings growth across a long tail of content driven in part by search and discovery advancements, and we made great progress on the technical underpinnings of our platform that drive genre expansion. We took safety a step further in just the last 2 months, rolling out facial age estimation across our platform, and we're unique in large platforms with over 100 million DAUs. In doing this, we completed the global rollout in January. Today, we continue to be bullish of our future. We're at about 3.4% of the global gaming content market. As you know, we're aiming for 10%. And internally, we have even more ambitious goals for the U.S. market.

Now diving into Q4, where we delivered another quarter of outperformance. Q4 revenue was $1.4 billion, up 43% year-on-year. And our Q4 bookings were at $2.2 billion, which is up 63% year-on-year. We saw strong growth across all regions. U.S. and Canada grew 41% year-on-year. APAC grew 96% year-on-year with a couple of key strength points, Japan at 60% year-on-year, India won 10% year-on-year and Indonesia at over 700% year-on-year. On DAUs in Q4, we saw growth at 69% year-on-year, really strong growth across all regions. And our engagement hours in Q4 of last year were at 35 billion hours. That's up over 88% year-on-year.

Quarterly -- monthly unique payers were nearly 37 million, which is close to a double from a year ago and up sequentially. And this was strong across the globe, U.S. and Canada up 34% year-on-year. And in Q4, DevEx was at $477 million, which was up 77 -- or 7-0% year-on-year. This increase in DevEx reflects our 8.5% increase rate in September. And for 2025, creators earned over $1.5 billion for the first time.

If we take a look at our top 1,000 creators, they earned an average of $1.3 million, which is up over 50% compared to a year ago. Our results reinforce our conviction in our long-term vision. We believe the future of gaming is part of a bigger vision around human co-experience. Our mission is to connect 1 billion people every day with optimism and civility, and we remain bullish about this vision.

In the shareholder letter, we outlined some areas of innovation and investment we believe will fuel growth in 2026 and beyond. First, we introduced the concept of novel game expansion, which is how we talk about expanding the genres and footprint to our older audiences. You'll note now that we're age checking all users who participate in communication on our platform, we've been able to find really a bigger growth opportunity in the 18-plus demographic than previously assumed. We estimate our 18 and over cohort is growing at over 50%, and this cohort monetizes 40% higher than younger cohorts. We're optimizing our platform technically to facilitate growth of high monetizing genres popular with older users such as shooters, RPGs and sports and racing. And we believe our technical opportunity on the platform is enormous.

When we look at the gaming industry as a whole, in a sense, there's enormous room for advancement right now. Typically, where we're going with our platform is a focus on vertical integration from cloud to engine to tooling to our clients on many devices to discovery, economy and safety as opposed to what typically and often happens in the gaming industry is in a bespoke way, putting together different cloud, different engine, different social communication.

We have an enormous amount of high fidelity AI training data that we'll share more with you, and you can watch on our main X feed as far as what we're doing with it. And we are pushing the notion for studios that the exact same gaming experience should run very well and at high performance on low-end 2 gigabyte Android. and at the same time, blossom into high res on a PC or console. And that includes running on every language or most languages with auto translation as well. Our stack from top to bottom is multiplayer, and we run that extremely efficiently in our own data centers. And of course, we auto scale. And we believe for cost, it's very important to run the majority of your world on your own bare metal, but be able to burst in the cloud.

Finally, we'll talk more about our advances in safety that are built in. A couple of highlights on critical innovations we're introducing to support genre expansion. We have now -- are in the middle of a full cloud rollout of native streaming of both 2D and 3D assets with various LODs. We're doing that automatically in our cloud. We have introduced SLIM, which is our nickname for dynamically compositing very complex assets in our cloud and delivering those at various levels of detail for performance on all devices. We are -- we've announced that we're bringing native server authority, which is very necessary and popular and competitive gaming. And we've rolled out custom matchmaking, which our creators can use to optimize either latency, age or social connection. In the first half of this year, stay tuned. We're doing a complete release of an expansion of our avatar system, including higher fidelity and more articulation.

And finally, really, we've seen over the last few years, the definition of what is a game expand. The more infra and platform technology we provide, the more we see experiences that aren't typically thought of as a game, for example, dress to impress and grow a garden. And we're sharing more and more the AI tech we're building on top of our enormous data to continue this expansion with the use of AI. This week, we launched 4D Generation, which allows experiences to include creation of new objects simply by users prompting and creating by AI. We shared earlier our vision on using our training data to create higher and higher quality NPCs. We're now working on photo upsampling in our cloud on both 2D and 3D for higher photorealism. And we shared really yesterday the work we're doing on our internal world model as a service and as a service that could be used both for creation by walking around and painting potential worlds as well as something we look forward to potentially integrating with Roblox moments.

In addition to this AI future, AI is really driving creation, safety, discovery, translation in addition to these potential areas of user growth on our platform. Every day, we capture roughly 30,000 years of human interaction data on Roblox in a PII and privacy compliant way. We're actively using this data to develop and train AI models that continue to bring our vision to life. I want to highlight that we're internally now running over 400 AI models. This includes, of course, Cube 3D, which has recently expanded to do 4D or functional interactive type simulation. We have our own facial tracking client side model that we use for avatar facial animation. Our voice safety model has been open sourced and is open source as part of the ROOST consortium. Our text filter is constantly updated and we believe one of the best, if not the best, in the world. Our text and voice translations are supporting published once and chat in multiple languages. And of course, we highlighted our internal world model team and our internal NPC efforts as well.

We see a future of creation in Roblox Studio that's enhanced not just by code gen and not just by experience gen, but ultimately, agentic iteration and testing in our cloud, once again powered by our 4D foundational and 3D foundational cube models.

Some other AI highlights, and then I'll turn it over to Naveen. In discovery, AI has been driving personalization. In Q4, AI drove a double-digit increase in the number of unique experiences that are surfaced in our Recommended For You algorithm. In 2025, on average, users engaged with over 24 unique experiences per month. This is up double digits from 2024. And I want to note we've done this all relatively transparently by sharing with our creators the factors we use for discovery.

Of course, critically in safety, we've been using AI not just for our voice model, but with state-of-the-art models like Roblox Sentinel and Roblox Guard. And I want to highlight in safety, the continued focus on our goal to lead the world in online gaming safety and civility. We shared a few months ago what we believe is the gold standard. We continue to innovate in this direction. Of course, no image sharing on Roblox, of course, monitoring text for critical harms. But in addition, moving to the estimation of age of everyone on our platform and the use in how we support chat.

In Q4, we started a global rollout of age verification for access to communication in Australia, New Zealand and Netherlands. I'm pleased to report adoption has been strong with approximately 60% of DAUs age checked in these markets. We completed our global rollout in early January of '26. And as of January 31, we've achieved 45% penetration of global DAUs. We're bullish about continued adoption with Australia and New Zealand and as a model. We are also rapidly enhancing our platform to make both age check adoption and to improve its reliability. And then kind of a fun full circle here that as part of our age estimation release, we're really going up and down our system. We've enhanced our matchmaking to cluster users based on their age and skill level. We're moving to continuous age estimation, which will use additional signals such as play patterns, the social graph economic activities to supplement facial age estimation. And over the coming months, we have more product launches, including always continuous refinement of our text and voice filters.

We're ambitious, and we believe these types of enhancements really give us the opportunity to enable even higher level of engagement than what we saw prior to our Age Check rollout. From a commercial and financial standpoint, our flywheel continues to accelerate. We believe having all ages on our platform is a long-term strategic opportunity that many other platforms are not confronting as they claim 13 and up users. We're seeing the growth we saw in 2025 in combination with fixed cost discipline to reinvest in our creators and our infrastructure, all really with an eye to fueling continued growth and long-term margin expansion. We're excited about the innovations we're developing and executing across all areas of our platform, which we believe will ensure our ability to continue to deliver on our long-term vision and deliver growth over a multiyear period.

With that, I'll turn it over to Naveen.

Thank you, Dave, and good afternoon, everybody. Dave obviously shared a lot of the highlights from a spectacular 2025 and a strong end of the year in Q4. I'm going to share a few reflections on the business based on what we've observed over the last few months and then also share some additional color on our expectations for 2026. We have more conviction than ever in the ability for Roblox to grow in excess of 20%. Our platform is healthier than ever. In fact, if you look at Q4, we saw very strong growth rates, 63% growth in bookings without the benefit of a big new viral hit. Our Age Check data, as Dave mentioned, showed that we have an even larger opportunity to grow with older audiences than previously imagined. That's especially true in the United States, where the 18 and over cohort on our platform is growing at more than 50%.

We have an international business that is still relatively nascent and growing at close to triple-digit rates. And the pace of innovation in AI is a tremendous accelerant and has the potential to help us grow beyond gaming as we develop even deeper connections in communication, entertainment and commerce. In addition to the tailwinds on the top line, all the ingredients for long-term margin expansion are in place. Now that margin expansion may not be linear, as you can see with investments that we're making in 2026. But in the long run, we expect to capture operating leverage from COGS, from infrastructure and our fixed costs. So as I said, our conviction in the ability to drive very healthy top line growth and bottom line margin expansion continues to grow.

Now at the same time, we've learned that it's difficult to predict exactly where this business will land 12 months out. I mean if you look back at 2025, when Roblox set guidance, Steal a Brainrot and Grow a Garden had not even launched. And that's created a situation where the company has had to provide relatively conservative guidance. I don't think that's helpful to investors, and it's certainly not helpful to day-to-day operation of our business. So we're going to get out of that cycle. We're going to give everyone a long runway. We're providing detailed guidance for 2026. But as we get into 2027, you'll see us starting to guide one quarter at a time.

So let's talk about 2026. We're expecting bookings growth of 22% to 26%. Those estimates are informed by the quality of the users that we saw come to our platform in 2025. It's informed by recent content trends that we've seen in early 2026. It reflects our confidence in the adoption of our age checking technology. And a number of things that we have planned in our road map related to our economy, discovery capabilities and many other features. Now importantly, our bookings guidance does not assume because we wouldn't be able to predict it, another viral hit of the magnitude of a Grow Garden or a Steal a Brainrot.

And when it comes to margins, we're expecting at the high end of our bookings guidance, margins to be relatively flat year-over-year. At the low end of bookings, we are estimating a slight year-over-year decline in margin. That's driven by the increase in the DevEx rate that we announced last year, and we'll see a full year impact of that in 2026. It incorporates investments that we've talked about related to continued growth in users and engagement and also AI workloads. And we're also planning to invest more aggressively in safety marketing to better educate our users, parents and other constituents about everything we're doing to ensure that Roblox remains a leader in online safety. And we are funding a decent chunk of those investments through operating leverage on COGS and fixed costs.

Our guidance also implies another year of strong free cash flow growth. In fact, at the midpoint of our guidance range, we're estimating 26% year-over-year growth in free cash flow. That includes a slight uptick in CapEx spend relative to last year as we are continuing to land GPUs in our own data centers and also navigating the recent inflation in memory prices.

So overall, we see 2026 as another year of strong growth on top of a spectacular 2025. We're making investments in our creators, in our infrastructure, in our safety, all of which sets us up for future shareholder value creation in the form of long-term growth and margin expansion.

With that, we'll open the line for questions.

[Operator Instructions]

We'll take our first question from the line of Ken Gawrelski with Wells Fargo.

Two, if I may. First, maybe, one, could you talk about -- could you give a little bit more color on the elements that inform your outlook for the bookings outlook for the year. I think it was -- it certainly has been a bit better than feared. Could you talk about -- I know you're saying that you don't assume any big viral hit. But like how much visibility do you have kind of beyond the first quarter into content schedules and releases? And kind of what informs your conviction in that guidance?

And then the second one, maybe just could you elaborate a little bit more on the -- you said the increased -- the better opportunity to target the 18 plus. And maybe within that, kind of if you could please talk a little bit about the recent developments with Genie and AI gaming. Just how does that inform your view and help hurt -- how are you going to use AI to make the platform better for developers?

It's Dave. I'll kick off a little on the first question, pass it over to Naveen and then come back on the second question. We have a lot of internal leading indicators we can see that somewhat correlate to the health of our system. And part of that is content diversity, content distribution, content velocity and types of content that are hitting different age ranges and genres. And we see a lot of health in that. So from just a high-level predictive view, that is one confidence inspiring thing we would see.

I'll hand it over to Naveen on more of the modeling and the makeup of our bookings forecast.

Yes. Thanks, Dave. I mean I'll really kind of put a finer point on what Dave said and then a little bit of what I said in my opening remarks. The content dynamic that Dave mentioned, I mean, we saw both in Q4 sort of an increasing diversity of content. We noted in our shareholder letter that experiences outside of our top 10 are growing at an even faster rate than they were in Q3, and that's true both with respect to engagement and bookings. So that's something we really like to see just spreading the growth around, if you will.

And we've continued to see similar trends in early 2026. And we like both the diversity of content, but also the freshness of the content. So there have been even not necessarily something as big as a Grow a Garden or Steal a Brainrot, but there have been new titles come in that have grown in a healthy way, which gives us a lot of confidence in the power of the platform.

I also mentioned the quality of the users that we saw in 2025. There was some uncertainty in that period of very high growth. Are these sort of low-calorie users that are going to come in and disappear. But when we look at what has happened on the platform, the behavior of the new users that came to Roblox in Q3, Q4, they largely look like our core users. And that's true when you think about how they engage, how they spend, how they retain. So that gives us confidence in how to think about the business over the next few quarters.

All that being said, there is still uncertainty trying to predict exactly where things are going to land 12 months out. That's the reason we don't think annual guidance is the right approach long term. We are using a slightly wider guidance range for 2026 than what we've used in the past, which kind of reflects some of that uncertainty. And then importantly, as I said, we are not building into our guidance an assumption of a massive viral hit the size of a Grow a Garden or a Steal a Brainrot. We're optimistic that things like that will happen again. But since we can't predict them, they're not built into our guidance.

Dave, do you take the second question?

And then just on the AI future, and I highlighted a little that we're optimistic we're going to see an expansion of the definition of really what is gaming and AI is going to power that. I'll step back though to our mission. And our mission is to connect 1 billion people every day with optimism and civility. And I'll highlight the word connect, which means bring multiple people together to play, to learn, to work. And for the last -- really since we've gone public, we've been sharing the visionary spec we've been building, which is many, many people coming together in physically simulated environments within games to play. We've started adding the notion of adding both NPCs as well as people to those environments, making those environments more and more photorealistic. And finally, making those environments one in which in real time, people can create, modify and change the environment.

So we're staying with that spec -- that spec. And we're -- as we showed on X yesterday, starting to use AI up and down the stack, up sampling 3D to bring things more to life, training NPCs with the 13 billion hours of not just video data, but intrinsic 3D world data. And we highlighted yesterday our internal world model team and what they've done using not just video data, but internal Roblox data to build internal world models that we think we can use both for creation and whatever.

I will highlight one thing, and that's very important to the spec we're designing. We're building multiplayer platform technology. We're building stuff that brings people together. And a lot of the current work that you see out there is operating in video latent space rather than synchronize 3D multiplayer cloud space. I would just cautiously highlight investors to understand the difference of that. You can read our recent tweet on X. That said, we continue to be optimistic about hybridizing and putting together many AI components to build the stack we're talking about.

Your next question comes from the line of Eric Sheridan with Goldman Sachs.

I appreciate all the disclosure in the materials and especially the shift around the guidance philosophy. Dave, I have one for you. The journey you've been on with respect to discovery on the platform over the last 12 months, what have been the key lessons you've learned about the opportunity that sits in front of you based on what you've learned about the evolution of discovery? And how does that align with your strategic priorities either for the platform or individual products as you look out over the next couple of years?

Discovery is a really hard problem to do right. And we believe being transparent in it is very important. We're as much as possible trying to optimize discovery not for short-term gain, but for long-term gain. And that's both long-term gain in user engagement as well as long-term gain in platform health and creator health. And any efforts to optimize discovery in the short term may not be optimal for long-term enterprise value and the health of our system.

So figuring out how to personalize discovery for every user in a way that connects great users with great content in the long term is what we've been focusing on. What we've seen as a result, as I mentioned, is more good content, reaching more great users, an increased robustness in the diversity of our content mix in a way that we believe is very, very healthy and also some future opportunities. So stay tuned with moments. We believe for many users, browsing in-game experiences will more and more be a complement to our discovery. So it's a long-term journey, and we'll keep getting better at it.

Our next question comes from the line of Matthew Cost with Morgan Stanley.

I have 2. Maybe one for Naveen to start. I think that the gross margins that you showed in the quarter were, I think, the second strongest that we can see going back even to 2020, I think it was the only time you did better. When we think about that improvement to gross margins, is that a function in the fourth quarter of a shift towards direct payments or any other moving pieces? And could you just give us an update on kind of the direct payment initiative?

And then a second one for Dave, if I could. I think the detail on the blog post that you've shared so far on this call about the work you're doing with AI and world models is incredibly helpful. And I want to put a finer point on kind of the concerns that we've seen coming up in the market over the past week about the potential for disruption to your business from other advances from -- with AI coming from other companies out there in the space. And I wonder if you could respond to those concerns and what you see as the differentiating factors that protect Roblox. I think the commentary you made a moment ago about multiplayer versus non was really helpful. But if you could expand on that.

Yes. Matt, thanks for the question. I'd really point to 2 things in relation to your question on gross margin in the quarter. One, as you hypothesize, we did get some tailwind from COGS. I think as we've spoken about in the past, we are doing a number of things in the product to try to steer the purchase of Roblox to lower-cost platforms. And that honestly performed better than we anticipated in Q4, and that was very helpful from a margin perspective. We -- the other source of margin expansion in the quarter was really bookings growth. Bookings came in also better than expected, which gave us powerful leverage against fixed costs.

When we look forward, and I think I spoke about this in relation to 2026 margins, we do continue to expect improvement in COGS rates as we are able to shift more and more of the business to lower-cost platforms. That's not going to be necessarily linear. There will be places and points in time where we're able to be more or less aggressive on that. But in the long run, we do see that as a source of additional margin expansion in the business.

Dave, on AI?

Yes. I would -- this is a great question. I would flip it and share how we think about it internally which is an opportunity for disruption in the opposite direction, which is an opportunity for the expansion of the Roblox vision beyond gaming into the future mix of what is entertainment and where does gaming end and where do other points begin. 10 years ago, a lot of people in the market used to talk about how is video going to get interactive, and there's been a lot of experiments on that, that have been very difficult. But the other direction to think about is how does gaming expand and become part of entertainment.

So I would say stay tuned on that. We feel what we're building, which is multiplayer technology that runs in the cloud that more and more can load instantly that more and more can be consumed in smaller bite-sized nuggets will more and more start to blur those 2 visions. World models are interesting. I would say, not initially in many of the ways that I think many people think, but we do think they have an opportunity for walking around and painting a world, for example, they have a really interesting opportunity to think about where does video end and snapshots end and interactivity begin. And so we have developed our own models there, but the core of our technology will continue to be the very difficult problem of 3D cloud synchronization and building communication type technology.

Our next question comes from the line of Omar Dessouky with Bank of America.

I wanted to ask more of a financial question. Specifically, how do you think about dilution in stock-based compensation in 2026 with your stock price significantly down compared to last year? Could you explain how having that stock-based comp flexibility is helping you make long-term strategic investments? For example, what investments besides headcount have a higher ROI that you would use that cash for?

I'll go first, and then I'll hand it over to Naveen. Internally, we're always running a multiyear model on stock-based comp and dilution at a very, very wide range of stock prices and running the business in a way we feel comfortable. So we do think about these things and model many dimensions. I'll hand it over to Naveen.

Yes. Not a ton to add to that, Omar. But number one, I would say we look at this on a pretty long-term horizon. I mean that both with respect to the share price and dilution. I mean, sure dilution at various points in time might spike just because of what's going on with the stock price. But ultimately, we think we're going to create shareholder value, and that will cause dilution to come down over time. If you look at what's happened over the last few years, that's definitely been the case. So something we pay attention to, but we're much more focused on the operating results of the business because dilution is going to get solved by that.

Our next question comes from the line of Brian Pitz with BMO Capital Markets.

Maybe a quick update on your ramping advertising ambitions and how you're thinking about the potential growth contribution from advertising in 2026? And then any additional detail about the age verification rollout, which maybe was not as smooth as you hoped. Can you comment on specific challenges and adjustments the team has made to ensure a better transition?

I'll go first. We sure don't think about it that way. We're very excited and proud of the way our age verification rollout has gone, and we're very optimistic that the result of it has been expanded thinking within our team on long term how to be unique in being a platform that can have all ages on the platform, can monitor and help how communication happens on the platform. I'll say that we gave our internal teams an ambitious goal of rolling this out eventually with no friction. And I would say, by doing this, we found so many other opportunities for optimization that I'm very pleased and happy with the way the rollout has gone.

And then on advertising, we expect our advertising business to show some pretty healthy growth in 2026. That being said, it is still modest. It's not a major contributor to the top line. As you've heard me say in the past, it is going to take some time to grow. I think the long-term opportunity is very exciting given the scale and engagement of our platform, but it is going to take some time, and we are executing carefully with respect to building out those products, the integration of the technology in our platform, working with our creators to expand the amount of inventory that is available, and we're going to continue to be very methodical about that to make sure we're building the business in the right way.

Our next question comes from the line of Cory Carpenter with JPMorgan.

I had 2. Maybe first, there's been some reports of interest in maybe building your presence in China. Anything you could comment on there and the type of opportunity that you see? And then just on the age of users, you framed it as the large opportunity given in novel game experiences given more younger users than previously reported. The half glass empty view of that, of course, could be that younger users have been tougher to age up on the platform than you expected. So what's giving you the confidence to invest there that you can age up more with users?

I'll go first on China. We continue to be a great partner with Tencent, and we continue to see a huge opportunity in China. We've revamped the way we look at China. If we were to go live in China, we would do it in an air-gapped way. And we think our infrastructure is built and designed in a really unique way that we can abstract and deploy it in multiple places. On our -- a couple of things about age checking and getting detail on estimated age. The first thing is it highlights the level of cultural phenomena that Roblox has become. And so yes, age checks slightly younger than self-reported. But if anything, it highlights the success.

I look to a couple of things. First, over 50% growth year-on-year, 18 plus. Two, the platform and technical advantages we've used to get to where we are in under 18 are exactly the same 18 plus. vertical integration all the way from cloud to apps to discovery to social graph and beyond, we believe right the same. And ultimately, the tech as well, supporting more and more realistic experiences. So I continue to be absolutely bullish on our 18 and up opportunity.

Our next question comes from the line of Shweta Khajuria with Wolfe Research.

This is Andrew on for Shweta. Just kind of one on the age check rollout. You talked a lot about the penetration that seems to be going well. But have you noticed any change in behavior or engagement levels for those who have completed age check versus not yet? And then maybe when you think about the derivative impacts of age check, is it possible to think about how the older cohorts may be viewing this as a quality of life update that might be contributing to the engagement levels that you're seeing?

I'll give one example on why we're excited about this. And that is the more we get into age check interacting with communication, the more we can more accurately match make different age bands together. That's one of the factors that makes me so optimistic is that age banding our matchmaking in ways that brings the average older user together as well as the average 15-year-old together, we believe can be a long-term growth aspect. So we think this is going to become -- it's why we call it the gold standard actually. And what we've seen after we did this is another very large gaming company announced they're going to do it, a communication platform announced they're going to do it. So we just see this as ultimately the way the world is going to work. We're proud to be one of the first big platforms to do it.

Our final question comes from the line of James Heaney with Jefferies.

Yes. Naveen, if we look at the Q1 bookings guide, it looks like a sequential decline greater than 20%, which I think would be one of your bigger sequential slowdowns. Is there anything you're calling out there specifically? Or is it just kind of the conservatism that you called out? Anything specific and even maybe what you're seeing so far in Q1?

Yes. I mean, I guess I look at it pretty differently. This is a quarter where we do not have, at this point, a big new viral hit. And so to put up 40% to 44% top line growth in the absence of that, I think we should all be very, very excited about what that says about the health of the platform. We said last quarter that as we started to move past this period of time with huge viral hits that growth would slow. So I don't think anyone should be surprised by the sequential trend there. If anything, I think we should be really excited about what we're seeing early in the year. I'm going to turn it over to Dave to close out.

Thank you all for joining us on our call today, and we appreciate the really interesting and thoughtful questions. Thank you all, and we look forward to seeing you in a quarter.

This concludes today's conference call. Thank you for your participation. You may now disconnect.

Good morning. My name is Kate. Welcome to Roblox's Third Quarter 2025 Earnings Conference Call. Stephanie Notaney, you may now begin your conference.

Thanks, Kate. Good morning, everyone. Thank you for joining our Q&A session to discuss Roblox's Q3 2025 results. With me today is Roblox's Co-Founder and CEO, David Bizzuki; and our Chief Financial Officer, Naveen Chopra. Our shareholder letter, SEC filings, supplemental slides and a replay of today's call can be found on our Investor Relations website. Our commentary today may include forward-looking statements, which are subject to risks, uncertainties and assumptions that could cause actual results to differ materially from those described in our forward-looking statements.

A description of these risks, uncertainties and assumptions are included in our SEC filings, including our most recent reports on Form 10-K and Form 10-Q. You should not rely on our forward-looking statements as predictions of future events. We disclaim any obligation to update these statements, except as required by law. During this call, we will also discuss certain non-GAAP financial measures. Reconciliations between GAAP and non-GAAP metrics can be found in our shareholder letter and supplemental slides.

With that, I'll turn it over to Dave.

Thank you, Stephanie. Good morning, and thank you for joining us today. A year ago, at RDC, we shared the goal of capturing 10% of the global gaming content market we'd like to report that we've made tremendous progress. We estimate now that 3.2% of global gaming bookings or revenue is going through Roblox, and that's up from 2.3% last year. Our platform and our creator ecosystem is healthier than it has ever been. Similarly to Q2, in Q3, we saw strength both across the platform with both existing and new experiences on the platform. The number of experiences that have more than 10 million daily active users on the platform at some point during Q3 2025 hit 7, and that includes a lot of experiences we're all familiar with as well as 5 that were created in the last 12 months. That's Grow a Garden, Steal a Brainrod, Brookhaven, 99 nights in the forest, PlantsF Brainrod, Ink Game and Block roots. We believe the success on our platform continues to be driven by raw platform capabilities and performance, continued improvements in discovery, continued improvements in our virtual economy, and complementing that, the investments we've made in infrastructure, which supported multiple groundbreaking records over the last quarter, including a 45 million peak concurrency during the weekend in August. Let's get into the financial results.

In Q3, our DAUs hit 151.5 million. That's up 70% year-on-year. This was growth really across all regions, including U.S. and Canada, 32% up year-on-year; APAC up 108% year-on-year. And importantly, we see continued evolution of our age demo with 13 and over DAUs growing 89% year-on-year. Right now, 2/3 of total DAUs are 13 and up. Hours had similar strength, hitting 39.6 billion hours of engagement in Q3. That's up 91% year-on-year. Strength across all regions, U.S. and Canada, up 47% year-on-year.

APAC hours of engagement up 127% year-on-year. commensurate growth with 13 and up 107% year-on-year, 68% of our total hours are 13 and up. Q3 revenue was $1.36 billion, up 48% year-on-year. Q3 bookings was $1.92 billion, up 70% year-on-year. Once again, strong growth across regions, U.S. and Canada, bookings up 50% year-on-year. APAC bookings up 110% year-on-year.

Some highlights. Japan, 125%; India, 46%. Indonesia bookings up 804% year-on-year. Our monthly unique payers continues to be strong at 35.8 million, up 88% year-on-year. And in Q3 through DevEx, we hit $427.9 million of DevEx, up 85%, a new record. I want to highlight that DevEx, which is what our creators are earning on the platform has grown 250% from the same period 2 years ago. We continue to have strong conviction for our long-term vision, and we'll continue to be more diligent about investing in this innovation to support long-term genre expansion and growth.

We've talked about genre expansion a lot over the last year. I want to do a few highlights here. We have a lot of technology that showed at RDC this year that's rolling out, including server authority and custom matchmaking, which will make Roblox more resilient and powerful for competitive genres like shooters, sports and racing. We've already showcased Avatar enhancement technology that is in the pipeline that we believe is going to really expand the look and feel of Roblox to higher fidelity and more lifelike avatars and continued focus in raw performance on the platform, including tech like Harmony and Slim, which we believe is part of the future of high fidelity gaming and the ability of our creators to make experiences that can run both on low-end 2-gig Android as well as really nice high-end gaming to PCs.

On Discovery, we've continued really a commitment to transparency that we also believe is good for the company. And we shared the notion that we're sharing our discovery signals with our creators, making those transparent in their analytics dashboard. And we've highlighted that we're using play-through rates, 7-day play days per user, 7-day intentional co-play days. So our creators can see exactly what we're using to make recommendations. We continue to see Discovery highlight new hits.

For example, in Indonesia, Fish it has become extremely popular, and we believe our Discovery system has helped promote that. I would also highlight our content ecosystem continues to be strong with what have now become perennial successes like Brookhaven Ad.me and Blocruits, which all continue to draw strong engagement. And some of our hits from last year like rivals and dress to Impress continue to launch active and successful updates. Really important at RDC, we announced an 8.5% increase to the DevEx rate.

Creator earnings surpassed $1 billion in the first 9 months of 2025. Also in support of our creator ecosystem, we've launched our IP platform, which we believe really is the future way to allow IP holders and creators to connect without all of the complexity of handcrafted individual one-on-one contractors or contracts. And I want to highlight IP owners like [indiscernible] and [indiscernible] have joined that platform. And we recently announced and launched Roblox Moments which we believe is an additional innovative discovery surface for our creators. Now in addition to all of this, I want to make a couple of highlights on some of the areas of tech we've been really working on before I hand over to Naveen.

First, really safety, which has always been a top priority for us and foundational to everything we do at Roblox. Just yesterday, we announced a partnership with the AGA, the Attorney Generals Alliance on a child safety coalition. Stay tuned with this. We believe there's a wonderful opportunity to share and develop what we believe is going to be the industry standard in communication for social and mobile apps, including our commitment by the end of this year to use AI-based facial age estimation to estimate the age of everyone on our platform and to use that to gate who uses communication technology and help root who can communicate with who, even in addition to what we already do, which is filtering of all text and no image sharing on Roblox. We believe as new technology rolls out, it allows us to harvest and use this technology for continued advancements in safety and civility.

We've also released over 100 innovations this year in our safety and civility group, including an announcement that we're going to be adopting IARC and International Age Rating Coalition rating over the next few quarters. And we've raised our minimum age for restricted content to really the global standard of 18 years old. I want to highlight, we do run stricter than typical industry policies on Roblox. We believe it's an essential strategic investment to how we run the company. And as we roll out facial age estimation, we really do believe this is going to add long-term value creation for shareholders, even if there are any short-term headwinds from that rollout.

On the AI side, we are now up to running over 400 AI systems inside the company. These are core within safety, discovery and creation, highlighting a few things that we've shipped so far. Our Cube 3D model, which we shipped earlier, is really going to come to life over the next 1 to 2 quarters as this goes live in multiplayer mode for everyone on the platform. We've also open sourced the Studio NCP server, which is making Roblox Studio the ability to integrate as both a client and a server in complex AI-based workflows. Behind the scenes, our safety model for PII continues to get better all the time. And we shipped RoGuard for our safety of our LLM text-gen tech in game. Stay tuned for a lot more generative AI.

I want to highlight that in addition to AI for safety and discovery, there's been a lot of chat about how much training data various people have access to. Within Roblox every day, we are moving forward to capturing, which is literally over 30,000 years of human interaction data and doing this in a PII-compliant way. This is unique data, data we have no intent of ever getting outside of our walls or selling that can really be used as we start to roll out our future vision of allowing people to play both with others as well as with NPCs and supporting unlimited creation in Roblox, not just for our creators, but for everyone. And that can mean per object with clothing, per the world you're in, per game creation and ultimately, with friends in real-time well playing.

Finally, 45 million concurrent users is a really big number. And we supported this really moving towards what we think is more and more an optimal mix of our own data centers, both core data centers, edge data centers and GPU installations on our own bare metal with bursting with our cloud partners. And we did this in really a wonderful way by bursting over weekends for several hour slots when we hit peak numbers. We're going to continue to go down this route.

We're investing more and more in our own bare metal for scale. Core data centers give us load balancing and efficiency. Everyone in Brazil is happy because we added a new edge data center there to reduce latency. We're going to continue doing that. And we are more and more now building out our own native bare metal GPU capability, but coupling with our cloud partners. We're going to make sizable improvements here. The big improvements we've seen in cost to serve may be harder to realize for the next few quarters, but this is all consistent with our long-term focus. With all of this excitement about innovation, we, of course, we have a lot of initiatives we work on every day to enhance our platform. We focus on the details.

And with that, we're going to hand it -- I'm going to hand it over to Naveend to complement my introduction.

Great. Thanks, Dave. Good morning, everybody. I'm going to try to keep my comments relatively brief so we can get to some questions. As Dave noted, the tremendous growth that we saw in the quarter was driven by this combination of big viral hits and underlying platform growth. And so with that in mind, I want to just share a few observations regarding overall platform health, if you will. Last quarter, I highlighted the engagement growth that we were seeing in experiences outside of our top 10. Well, this quarter, the growth in engagement for these experiences, again, outside of the top 10 accelerated even further from 47% in Q2 to 58% in Q3. That's the engagement. And then on the monetization side, spending in that cohort of experiences remained north of 40% -- excuse me, spending growth remained north of 40%. We also saw healthy growth in bookings per daily active user.

That was true in every region other than APAC, which I call out because very importantly, in APAC, the trend really is a function of mix shift at the country level. Payers, as Dave highlighted, grew 88% year-on-year, which is a very healthy growth rate in and of itself. But notably, it's higher than the rate of user growth, which was 70%. And we believe that this dynamic is at least in part caused by changes that we've made in our economy.

Remember, we launched regional pricing for marketplace items back in June. And we think that, that helped drive higher payer penetration in markets like Southeast Asia. We did see a decline in bookings per payer on a blended basis, but similar to my comments on bookings per DAU, it's important to understand that this was really driven by geographic mix shift. So platform evolution looks healthy, and that really increases our conviction in our long-term goals. But as Dave pointed out, also underscores the importance of continuing to invest to ensure that we can deliver healthy growth over a multiyear period of time. So what does that look like financially?

Well, on the top line, we see momentum in the business to continue to deliver healthy double-digit bookings growth. We think that will be aided by the launch of a number of the key technologies Dave touched on that will roll out in late Q1 and early Q2 of 2026. Many of these technologies are designed to enable genre expansion on our platform. You saw some of those demo at RDC.

From a reported growth perspective in '26, we think the adoption of these technologies will be a factor, meaning how quickly developers adopt some of these new capabilities will influence the growth rate, particularly given the tough compares that we know we'll have coming out of 2025. We're also conscious of the fact that the new safety policies we're rolling out consistent with our commitment to being the gold standard for safety may cause some short-term friction to engagement and bookings.

But ultimately, we think those are a magnifier of long-term growth. And then on the expense side of the equation, we are envisioning more investments in DevEx, in infrastructure and people to support our goals around safety and genre expansion. That's the reason you're not seeing year-over-year margin expansion in Q4 based on our guidance.

And it's the reason that we expect margins to decline slightly in 2026, given the combined impact of a full year of higher DevEx rates, limited cost to serve improvements around infrastructure and safety and higher growth rates in our comp and ben expense lines. There'll also be incremental CapEx starting in Q4 2025, that's incorporated in the guidance that we shared. And I would tell you to expect similar levels of CapEx in 2026, which, of course, implies that CapEx intensity next year should be lower than 2025 and but still somewhat elevated relative to 2024. So I realize probably getting a little bit into the weeds there, but let me step back for a second. I think the key takeaways here are we are way ahead of our long-term growth plans. And in fact, the reality situation is that bookings have grown faster than our ability to deploy the appropriate growth investments. And that means you're going to see some slight margin compression as we catch up over the next few quarters.

But I think those investments really should give everyone even more confidence in our ability to continue to deliver sustainable long-term growth. So hopefully, that's some helpful color both on what we're seeing and what we're expecting.

And with that, I think we'll open the line for questions.

[Operator Instructions] We'll first go through Matthew Cost at Morgan Stanley.

Dave, I want to start on AI. I mean, just when I look at your infrastructure plans, clearly, you're very excited about putting GPUs and more data center power behind what you're doing tie that back to what the user experience or the games on roadblocks are going to look like in this world that you're building towards. When you think about what Cube is capable of, what will real-time content generation and what you're calling 4D content creation mean for robots experiences and for engagement over the next couple of years? And then I have one for Navin as a follow-up.

Matthew, great question. And this is a fun one to answer. It also highlights the just enormous future technical innovation. We have rate looking in front of us. we have shared publicly RDC and others really a vision of what the ultimate spec for roadblock should be. And that ultimate spec should be creator decides everything from fun animate look and feel all the way up to photorealism, anywhere from 1 player to 100,000 people out of concert simultaneously. .

The ability to use AI to do real-time modification of that world, everything from a piece of clothing, to really 100,000 people with Dungeon master modifying the whole environment in real time. A mix of true human players, NPC players all together doing this with a very tight eye towards efficiency and cost because we're a premium platform. ultimately, doing this on everything from a 2-gig gram Android device to a high-end gaming PC. So that is that is an amazingly complex spec. What you're going to see coming out soon with Cube for the is real-time generation in experience in multiplayer, not just the static objects but of complex objects, vehicles, weapons, other types of things that user can interact with in multiplayer. -- more of that to come, more AI supporting the stack everywhere, but we're marching as quickly as we can to that really big visionary spec.

Some of the gameplay that we will see, we do not know what it is. We've seen that historically on roadblocks more and more as we both cover new and existing genres in a predictable way. along with new types of games, I would highlight Dress to impress was an example a year ago where the raw technology we were building to support more traditional gaming allowed the creation of a whole new genre. So stay tuned both for more coverage of existing genres at higher performance as well as as of now unthought of types of gameplay.

[Operator Instructions] We'll move next to Brian Pitz with BMO Capital Markets.

Dave, you mentioned increasing the DevEx rate by 8.5% putting more economics into creators hands. However, at the same time, if you look at competing UGC platforms like Fortnite, they're also offering very attractive economics to try to win over creators to build a mere UGC platform. How do you think about the need to continue driving economics towards creators to help send off any competition from other platforms that either are currently in development or could be coming in the future.

Great question. I do want to highlight, generally at roadblocks, we run the company by looking to the future and not looking over our shoulder I would say in this situation, there's one additional component to what is the DBX rate, and that is the ability of a new creator or an existing creator to make enormous economic returns. And so one needs to multiply the DevEx rate by the volume of users on the platform by the breadth of the creative tools by the velocity that new creators can make new experiences.

And in the end, that's how I believe and we believe creators analyze the situation, not simply by the DBX rate. We do believe for every incremental percentage within obviously our fiduciary duty and our balancing of earnings and Devex and all of our other costs that for every incremental percent. We do believe there can be some effect on the creators in making robots a more appealing platform. It ends up at the highest level making us think about our company, where we want our COGS to be as lean and as efficient as possible.

Our personnel cost to be as lean and efficient as possible. Our infra trust and safety costs to never compromise and at the same time, to be run as thoughtfully as possible -- and this has been one of the benefits of building out our own infrastructure. And then finally, being very thoughtful on how much we move back to our creator community relative to our own cash flow.

So I think it's much bigger than kind of this one-to-one comparison. And that's really long term, we want to move as much money in a prudent and thoughtful way while always being responsible for trust safety and our earnings to the creators.

We'll move next to Eric Sheridan with Goldman Sachs.

Can you share with us key learnings as older age cohorts continue to scale as a percentage of the mix in the business. And based on those learnings, how can you line up your investment priorities to sustain that growth and stimulate that aspect of mix in the years ahead?

I'll share some key learnings, and then I'll share how we are lining things up. I think one of the key learnings is roadblocks has a huge ability to virally attract new users to the platform with new hits. And we've seen that both with dress to and press. We've seen it with [indiscernible] where we really get user acquisition at enormous scale organically as word-of-mouth traverses these properties. We've also seen with new types of game play, once again, use those 2 as examples, that older players are excited and interested in that. .

If anything, as we look at the global gaming market space, which is estimated anywhere, I won't make the exact estimate, but numbers between $180 million and $200 billion we see all of the existing genres, and we've been able to align those genres with our technical road map. As we shared before, a lot of our technology, we believe, is going to support sports. A lot of our technology will support racing. Some of the new technology we're working on is going to make RPGs better.

Some of the technology we're working on now we believe we're going to see more and more Avatars on roadblocks that have a much bigger diversity just as we see in the gaming ecosystem as a whole relative to what we have on roadblocks. And so we can use that to guide our technical road map. But I would highlight we're not copying. We are envisioning satisfying those technical constraints while at the same time, building a platform where the exact same experience can run a 2-gram Android in a very difficult networking configuration and at the same time, look absolutely amazing on a high-end gaming PC, eliminating really kind of this void between mobile and desktop and console.

We also believe the future of platforms like roadblocks is much more cloud integrated such that AI and generative AI is always available in the experience for all creators. So we are able to align a bit both our own vision as well as what are all of the current genres in the gaming ecosystem and make sure we line up with that.

So that's -- we're optimistic there's a lot of growth ahead of us in some of these genres where we're not fully at 3% of the global gaming market right now.

We'll move next to Jason Bazinet with Citigroup.

I just had a question on the shareholder letter. I think in your '23 Investor Day, you laid out 20% plus bookings growth from '25 to '27. And in the shareholder letter, you acknowledge the great results you've had so far this year. But then you say as we look to next year, our long-term objectives have not changed. Is that essentially a soft way of saying that you think that the growth will be below 20% in '26? Is that what you're trying to say? Because I think that's what the market is trying to digest with the premarket exit in your stock?

Yes, Jason, it's Naveen. Thanks for asking to clarify that. Look, I think we are not providing any specific guidance about 2026 at this point in time. I think that would be premature. We need to land the plane on '25, and then I think we'll be able to dial in expectations for 2026 more specifically. I think what we're trying to highlight for people is that there are some important things to consider as we look at expectations for bookings growth next year. there will be tailwinds from the momentum that we are seeing in the platform today.

There will be tailwinds from a lot of the tech that's going to be hitting the platform in the first half of next year. But there could also be potential headwinds, obviously, from the tough comps and then potentially from some of the new safety policies that we are going to be rolling out. We don't think any of that changes where we expect this business to be over the next several years. But I think it is too early to put any specific numbers around '26 at this point in time.

We'll move next to Cory Carpenter with JPMorgan.

Maybe building on that, you did not mention advertising as a potential tailwind next year. So just curious what your early learnings have been in rewarded video. How big a push our priority do you plan to make that in 2026?

Yes. So I think consistent with some of the comments we shared recently around advertising. We remain very bullish about the long-term opportunities there. that we are cautious about the near term because we want to make sure that we get it right. And as you've heard at RDC, we are now rolling out rewarded video on sort of a limited basis. I think we pointed out in the shareholder letter, we now have over 140 creators onboarded.

What we have learned from that is that we want to be very thoughtful and diligent with those creators about how we integrate rewarded video to make sure that it works for them from an engagement and a monetization perspective, that it works for our users in terms of the experience, performance, et cetera, and that obviously, we are representing our advertisers in a high-quality way.

So there's still a lot for us to do on that front, and therefore, not something that we would call out as a major contributor in the short term. but something that is going to be a key part of our business as we progress over the next few years.

We'll move next to Ken Gawrelski with Wells Fargo.

Could you talk a little bit about the various genres? And how do you think about the the recommendation engine, which has been clearly very successful in kind of surfacing new hit games. Could you talk about how you're pushing more diversity of genres and to make sure that there's enough experiences out there. And is this a concern? Are you seeing -- are you worried at all a little bit about more concentration in certain types of genre games or in certain game mechanics?

Okay, great question. And part of my intro was to highlight that diversity on the platform with titles over 10 million DAUs at some point during the quarter and 5 of them new. One of the -- the way we have been thinking about Discovery, there's 2 big areas of it.

One is, we're not trying to optimize the short term. We are trying to optimize the long-term enterprise value of the company. And we're also trying to optimize ecosystem health in addition to what we put in front of users, which really does mean surfacing new creators and new genres side-by-side.

The second part on our discovery, we're moving more and more towards complete transparency of our discovery algorithm. We're sharing the signals that we believe point to long-term health of the ecosystem. And as you correctly noted, we believe part of that long-term health is through expansion in some genres, like RPGs, shooting, racing, battle, sports and other genres on the platform. The final thing I want to highlight on the way our Discovery system is working, is long term, there's really 3 prongs to discovery on roadblocks.

There's our own recommendation engine. It's coupled with on the homepage sponsored tiles which is paid discovered by our creators. And more and more new creators who are launching a new game or bursting are using that to complement our organic discovery. Finally, our top pick and curated sort is an additional way that we complement that with the vision on the platform. We're being very careful not to look backwards, not to burn in what has historically been big on road blocks.

But instead to look forward to what we believe will be new types of gaming on the platform. So I believe relative to a year or even 2 years ago, our discovery systems are working better than it ever has, and I think we have continued improvements coming. So stay tuned.

We'll move next to Benjamin Black with Deutsche Bank.

So just touching on the growing number of experiences that are eclipsing 10 million users. So with the growing data set that you have on user engagement, the AU behavior, does the Discovery model just get increasingly smarter at recommending content. So I guess the question is, are you starting to have a real data mode in discovery? And then Dave, over the past couple of quarters, we've seen a divergence in our engaged growth and bookings growth. Could you just help us understand -- so why is this happening is as easy as sort of a mix shift towards low monetizing experiences or low monetizing regions? And if so, how should the relationships between those to evolve from here?

This is a really good question. I want to put it under the umbrella of first privacy safety PII safety, adherence to all local laws and all of that? And second, I want to put it under the umbrella, but we're not under any financial pressure and have no intent to ever release any of this roadblocks data anywhere off the platform. Once we look into those 2 things, though, the data set is enormous. The data set is not simply who clicks where. The data set is real-time 3D Avatar interactions.

It's what areas of a certain experience are retaining more. It's what are Avatar is doing in any creators experience. It's the ability to analyze that and the ability to analyze the data that makes up a 3D experience. And in addition to maybe more traditional signals for discovery use 3D and time-based immersion signals to help predict what may be a good experience on our platform. that goes back to an interlock with what I shared, which is every day on roadblocks, there's over 30,000 years of potential 3D interaction,

Avatar training data available for us. So the answer is yes. We will mix more content understanding, understanding of what makes experiences interesting and fun into our discovery mechanism. And I'll kick it over to Naveen for the second part of that.

Yes. Thanks. Ben, your -- or at least the second part of your hypothesis was correct in terms of the dynamic between hours growth versus monetization growth. meaning it really is about geographic mix shift. You can see that if you look at -- in the supplementals, we had some disclosure around bookings per daily active user. I realize you were comparing monetization versus ours, but it's a very similar dynamic, i.e. If you look at this on a regional basis, very strong growth year-over-year. but the mix of hours is skewing towards regions that just do monetize at a lower level. And that's what you're seeing in the blended numbers. And as I pointed out in my remarks, even at the regional level, we see some geographic mix shift between specific countries that impacts the monetization at the entire region. So it really is all about growth in various regions.

We'll move next to Omar Dessouky with Bank of America.

So you called out the tough comp in 2016 and some of the dynamics there. I'm looking back at road blocks. A lot has changed, obviously, the company has almost doubled in size over the last 2 years. And when things change, you may consider different approaches and with '26 being such a tough comp year potentially, that it sounded like Dave said that the virality of your hits is the main engine for user growth. In a tough comp year, like 26, why wouldn't roadblocks consider advertising to kind of smooth out these troughs and peaks in in growth. That's my question.

Omar, great question. A couple of things to take a step back at big picture right now. we're simultaneously excited that we've nudged over 3% of the global gaming content market running on road blocks. And at the same time, there's almost 97% out there. So I continue to be enormously bullish as we roll out our tech and we expand genres, and we do this really in a new way for gaming complemented with AI.

The second thing I do want to highlight is that this quarter, next quarter, we're rolling out a lot of what we shared at RDC, which supports massive expansion of the way roadblocks works and the types of genres. You correctly note, robots got to where it is primarily based on viral growth, but we do buy traffic. I don't know if it shows up in our financial statements, but we very thoughtfully complement that with paid user acquisition. And what we are testing and starting to roll out is the notion that we can partner with the creators on our platform to help them supercharge their paid acquisition in partnership with us.

And there is a future opportunity for us to expand paid acquisition, where a creator may be a little hesitant to buy paid traffic and send that to roadblocks because we get some benefit as well. They land in our client. That user may pay other experiences. But when we start sharing that with creators there's an opportunity for both creators to expand paid acquisition and for us to support them and do it with them. So we already do some paid acquisition. There's an opportunity to make this a lot larger.

We always do this in a financially prudent way, which was appropriate return on ad spend in the right amount of time so that we're doing this incrementally. So yes, a great idea. We've got this in our sights to expand. And Naveen, I don't know if you want to complement that at all.

I think the only thing I would say is that in terms of overall scale of what we're spending on growth marketing it is still very modest, and we rely on the tremendous organic growth that the platform still has. As Dave said, there's a lot of market white space yet for us to tackle we added, I think, close to 40 million users relative to Q2.

So the organic growth engine is working really well, and that will still be, I think, the primary driver of growth. But there are some very interesting things that we can do, as Dave described, in conjunction with our debts to promote some of the content that is coming to the platform, which is consistent with our goals around genre expansion, content diversity, content velocity, et cetera. So looking forward to continuing to experiment with that.

Time for one last question from Clark Lampen with BTIG.

I'll try to make these quick ones. Dave, you talked about making the platform more attractive to the creator community over the last 2 quarters, we've obviously seen a really big spike in users. You've invested in developer exchange fees, tech, the platform. We obviously have a very good view into user growth, but a little bit less so on the developer side of the ecosystem, has debt growth essentially held serve with users in a way where we could think about that sort of growing proportionate to DAUs?

And maybe just take that a step further for Naveen, if we think about that and sort of the elasticity of growth and engagement, we've been through investment cycles in the past, this 1 feels a little bit different because it's AI-oriented. Is there a reason to think that like the yield on some of these investments, if we are seeing proportionate growth could be a little bit more immediate and is that reflected in some of the comments around '26?

Yes, I'll go first, and then I'll hand it over to Naveen on the yield side. At RDC, and I don't know if we published as part of this earnings, there's 1 spec or stat that we constantly share which is some averages around top 10, top 100, top 1,000 deaths and the year-on-year growth in bookings and engagement. And we continuously see, I would say, a flattening of that tail where the top 1,000 looks very, very healthy. So I think we're in a very healthy zone still. We continue to flatten that curve. We have many, many more creators making a living or making $1 million or making $10 million on the platform, and that all is continuing to grow. I believe we can extrapolate to the future, which is our vision of having 10% of that global gaming bookings running through our platform, and that gives a little bit of a perspective of what we would hope happens to that wide range of creators on the platform. Just one highlight on infrastructure. We continue to highlight the movement to supporting our own bare metal data centers, both for core edge and AI and complementing with Burst we're -- this has been part of the key to our efficiency, and I'll let Naveen speak on the yield of these investments as we get into that.

Yes. I think the way to think about, call it, the payback period on these investments is that there are a variety of things that we are doing that we will start to see the fruit relatively quickly. There are others that are more long term in nature. the AI stuff, in particular, there's a lot of AI investment that we've already made. You heard Dave talk about some of the benefits we're already seeing around discovery, economy, et cetera. So I view those as having relatively near-term payback. There are others that are going to be longer term. And quite frankly, it's going to take us a little while to roll out the CapEx to train and ultimately deploy these models to move them on to our own metal, as you heard Dave describe. So those are not necessarily going to have as immediate of a payback. But I think it's fair to say we're already getting benefit from a number of the investments that have been made on this front, call it, in 2025 and increasingly in '26 and beyond. I think that's our last question. So thank you all for joining. And then I will turn it back to Dave for any closing comments.

I just want to once again thank you all for your support and your insightful questions today. We look forward to continuing to innovate in our quest to have 10% of global gaming on the platform. Thank you all.

Thank you. And that concludes today's conference call. Thank you for your participation. You may now disconnect.

And now, please put your hands together for the Founder and CEO of Roblox, David Baszucki.

Oh my gosh, wow, it gets better every year. Last year, we were all here together. We shared an audacious goal of having 10% of all global gaming content running on our platform. And what this really means when you think about it is everyone in this room is part of that 10% of the global gaming space. We're well on our way. Last year, we were at 80 million daily active users. Today, we're at 112 million daily active users. That's not 20%, that's not 30%. That's 41% year-on-year growth.

And last year, unbelievably, our peak concurrency was at 11 million. Two weeks ago, thanks to all of you, we had 45 million people playing Roblox at the same time, what a crazy week. And if you notice, thank goodness for those Brazil servers, they are now on, finally.

Okay. So let's honor what we call the Deka-Mega DAU. For those of you that know your prefixes, that's those experiences that have hit 10 million daily active users per day. Last year, we had 3 of you. This year, we'd like to welcome all 9 of the Deka-Mega DAU. Thank you. But it's not just the Deka-Mega, it is all of you over -- not $1 million, $1 billion was earned by our community in the last 12 months. That is up 38% year-on-year. And we now have over 100 experiences generating more than $1 million a year. Who would have thought that? So thank you.

It's once again not the 9, it's not the top 1,000, it's all of you. Today, we have a big overview. We're going to dive into our vision for safety and civility, not just on Roblox, but we're going to share with you what we believe is going to become an industry standard. We're going to dive deep into tech, the 3D scale and how you can make better games, not with AI necessarily, but games that incorporate AI. So stay tuned on that.

We're going to share big innovation coming in search and discovery, a big innovation. We're going to dive into IP and content. And of course, we will discuss our economy. So first, let's dive into safety and civility. So first, we would -- we just want to acknowledge all the discussion over the last few months. Safety is critical really not just to me, but to all of Roblox. We really appreciate the open dialogue.

We are doing a ton. We counted the innovations we shipped in the last year. We shipped over 100 safety innovations and initiatives in the last year. I just want to go through a few of them, and then we're going to highlight what's coming.

Trusted connections with people you know and trust. We've open sourced a lot, our voice filters and our text filters. We've open sourced Sentinel, which is our early detection system for preventing child endangerment. We've expanded experience guidelines, and we're now requiring everything to be rated and we're showing age on every experience just to help people understand what we're already doing. And of course, we moved our most restricted age range from 17 to 18 plus.

For those who have been with us from the start, you may remember early on, we had a fair amount of flexibility with creativity. I don't know if any of you remember flush a noob down the toilet, that went straight to the top of the charts for those of you that were here early.

I do want to highlight, and I want to thank you that as our global audience is expanding, we are being a little more restrictive on policy issues. We appreciate those who are adjusting your content as we address them.

And I want to highlight what we shared this week, which we believe will become best-in-class in the industry. And that is we're going to announce and we have announced that we're going to expand age estimation to all Roblox users who access communication.

And rather than just relying on what someone types in, we're going to use AI facial age estimation and ID estimation to essentially superpower safety and civility on Roblox. We're also going to be very thoughtful about communication between miners and adults unless they know each other in the real world.

And I want to highlight, this is all in addition to everything we're already doing. No image sharing on our platform, world-class text and voice filters, but we do believe this will become the standard and the best practice for online destinations all around the world.

And you know our vision. We talk about it all the time. Vision, we want to be many times safer than the world's best amusement park for young people. And for older people who are ID verified, we want them to communicate freely and openly in, for example, a really edgy Western historical novel. We thank everyone in our community for the passion around this.

Now the more we work on Roblox, the more we actually feel the technology in the whole gaming market is literally in the stone age. And there's incredible room for technical innovation here. The industry has a long history of having studios try to glue together all kinds of stuff, engines, cloud providers, safety systems, payment systems and more. And from the start, for those of you that have been with us for a long time, you know we've had a very different vision.

Our vision for Roblox in the future has always been a unified, very tightly integrated system that delivers everything as one, not just in the client, but in the cloud and in all of our systems. And today, we're going to take a little look at where we're going with the vision.

But before we go anywhere with the vision, we need a spec. What are we trying to build? It's really pretty simple, really, pretty easy spec. 100,000 players simultaneously, imperceptible latency. When you watch your avatar, it's got to be a mirror of you. We want this in a photorealistic, physically simulated environment with people all around the world possibly speaking different languages. Of course, on a platform that is safe and civil, accessing any device, phone, tablet, computer, maybe TV someday, with instantaneous joins.

Your friend has a new game, you're playing instantly. You're not waiting for a download, where creators can push updates at any time. And with AI accelerating creativity in experience, not just for one or many players, it goes out saying that is ridiculous. That is a ridiculous spec. It's going to take a lot of work to get there, but that's kind of what makes our job fun at Roblox.

Now we're pushing closer to the spec with realism and scale with a lot of new technologies. First, confirming here today, server authority coming to Roblox native soon and quickly. So for Battle Royale, for competitive genres, Battle Royale shooters, fairness is crucial. Here's a prototype of server authority. We're going to be shipping by the end of this year. The view on the left is a point of view of something we see all the time, actors using client-side exploits. On the right, you're seeing the same experience from another user's point of view. You can see the exploiter with server authority, avatars are simulated on the server, aiming for 100% fair experience for everyone. The bad actor might see their hacks locally, but it's an illusion as we can see from the view on the right. We're opening up early access to server authority tomorrow.

We're also introducing today a group of technologies that are going to support massive high-fidelity worlds. We've been thinking about this. We've been talking about this internally for 5 to 10 years. We're starting to -- this is happening now.

First, our entire asset pipeline is being overhauled with on-demand cloud transcoders. This means we're going to generate multiple levels of detail for every single asset on Roblox automatically. This means every mesh, every texture, every animation, every sound is going to be transcoded in the cloud and stored at multiple levels of detail. And we're going to transcode assets like textures, so they download natively with whatever is best for the device of the user. And yes, this includes video assets as well.

Now in addition to cloud LoD, our Harmony system dynamically adjusts the load sequence of every asset LoD during a game join because we want to maximize the human perception of a fast join. And Harmony balances high LoDs near the player with lower res assets further away. The system on every client is watching CPU, GPU, RAM, bandwidth everywhere. And we're not stopping at cloud LoD or Harmony.

Someone just asked me yesterday, what about an avatar with hundreds of pieces of clothing, like what are we -- how are we going to solve that? Today, we're also introducing SLIM, which stands for scalable, lightweight interactive models. SLIM converts models into lightweight representations in the cloud automatically, so assets far away can be loaded with fewer meshes and textures. It's a compositor, dramatically improves performance and SLIM understands the complete context of your model.

So instead of just doing LoD for individual assets, SLIM can take a model all the way down to a single mesh and 0 client instances and anything in between. And of course, if we have LoD in the cloud and we have Harmony and SLIM, why wouldn't we do one last thing and that is offer almost unlimited asset resolution which we're going to do. And when we put it all together, it's absolutely magic. This is a real screen capture of a real Roblox place with a fly-through using SLIM and cloud LoD and Harmony altogether.

Assets close to the camera are rendering high res, assets further away are low res and ultimately, SLIM representations. And the detail you see far away are SLIM renderings with a single mesh and a single texture. We believe this is the future. So you can design and build at any resolution you want. It's all automatic in the cloud. Hopefully, millions of triangles someday. You can access and sign up for early access to SLIM today on Dev forum and early access starts Monday.

Now in addition to kind of scale, Avatar is very, very important for us because we're trying to capture the diversity of the human condition in games and movies everywhere. And for those of you that have been with us for a while, you know our vision is interchangeability. Anybody, any clothing, driven by any type of motion, can interact with any object in the world. This commitment to human expressiveness is why we're bringing native makeup to Roblox early next year.

Makeup is going to support layers of blush, lipstick, eye shadow, all blended with PBR, all blended with transparency. And makeup is going to work on any avatar face, including classic faces, and makeup is going to launch in studio beta in the next few weeks. We're also increasing the visual fidelity of our avatars. Our goal is to support rigs with 200 joints and beyond, all while keeping the interchangeability of all animation and all rigs. And of course, we are finally going to have native hands and fingers on Roblox, all thanks to our new movement system.

We're going to be bringing integrated kinematics with physical simulation on avatar joints. We're bringing crouch, sprint, straight, native. We're going to be introducing an animation compositor and inverse kinematics and an animation editor that's going to support inverse kinematics in our file format.

And when we put it all together, any avatar on Roblox will be able to interact with any object. This will include retargeting, so any avatar can pick up a weapon with both hands or maybe a dowsing rod all natively. And a big part of this design for those of you that have been watching us, for both cloud LoD, for Harmony, for SLIM, for unlimited res and for our Avatar system, is having anything run well on any device. And we've been saying it for a year. Our low-end target device is 2-gigabyte RAM Android phone. Here's the same fly-through we've been looking at running on the equivalent performance of a 2-gig RAM device. No other gaming system supports this range of integrated server client scalability.

So when we think about our creators and our users, we're always thinking about how can we build the most interesting creative games and experiences. How can our users, not just you, but our users become more creative. And 5 years ago, we used to talk about a fashion design where our users would create fashion. We're now starting to think about fashion design games and experiences that you build where users create fashion with generative AI. And so for 4 years now, we've been an AI shop. We're running over 400 AI models. It's just the tip of the iceberg. We've open sourced 4 of our models.

I run into start-up founders who are building video game companies who are starting to use our voice safety model, for example. But our long-term vision is really for AI to build and tune foundational models for use within Roblox based on the enormous data we have at Roblox, and we're rapidly moving to this. First, we want to support not just 3D but 4D generation. That means functional objects, physically simulated objects that can be created directly in an experience. This is different from any other texture mesh gaming platform. Roblox is about functional, physically simulated objects.

And for those who wish, we want to support not just model but true experience generation directly from prompts, whether in studio or within an experience, text prompt, image prompt, you name it, and this same technology can support up sampling. For example, you may want to design in blocking mode, but auto up sample to a photorealistic mid-evil mode kingdom. And ultimately, we're going to support what we call real-time dreaming. This is also referred to as procedural generation. It means human or an AI can alter the world in real time while people are in it.

And finally, of course, we want to support great NPCs, fully functional directly from a prompt, you design your NPC, you talk about what it looks like. And we're putting all of this together now with our toolkit for text gen, text to speech, speech to text and filtering.

So today, we're going to do a live demo. We're going to -- I have Karun Channa come out, Director of Product for AI. We're going to hop on stage, come on out, Karun, and we're going to do a little demo. So welcome.

Okay. So while you get fired up, Karun, we're wiring together text to speech, speech to text, text gen, 4D gen and...

Translation.

Translation. And so you're going to see interactive avatar with voice rather than text. We're not used to. And I think we're going to do some 4D gen in games. Take it away.

All right. Let's walk in. So what we have here is we build this game. Character over here is going to go out on a world exploration and this -- we have 2 AI NPCs here to train us out. So we're going to walk up probably unmute myself.

[Presentation]

So we're doing 4D gen right there in real time. And I want to highlight later in the show, we're going to give a little hint at the prompts that are driving these avatar, so stay tuned.

And I'm going to do one more.

[Presentation]

Now, are these hard-coded demo objects?

I don't think we want to try a watermelon gun. That's cool. The trainer kind of overheard the conversation and is able to kind of make a comment about it just because of the LLM.

I've never seen anything -- the kind of overheard the conversation. to make a comment about it just because of the LLM.

[Presentation]

So we saw 4, and we're going to see the fifth one now. We're going to Spain, I want to learn how to speak something in Spanish.

[Presentation]

I hope that was right...

That immediately made me think we're going to see some language learning apps on mobile.

Now the generations are ready. So that is my little bit on a gun. Let's see what a watermelon shooter looks like.

So highlighting this is a functional object with embedded code, not just a good-looking object.

That's right.

And we're going to see if it really functions.

Right. Here we go.

It does. Okay.

You see the watermelon. We're going to do the target practice, boom.

[Presentation]

Same thing. Not hardwired.

Not hardwired.

Just the magic carpet 4D generated.

And I think the carpet looks pretty nice, don't you think?

See if it flies.

There you go. Got to avoid that tree. And that's all we got.

All right. So thanks, thanks a lot, Karun. So we've already seen launch text gen and speech-to-text. Speech-to-text is coming to beta before the end of the year. By the end of the year, functional 4D generation is coming. Next year, we're going to roll out real-time voice chat translation. So stay tuned.

We've been focusing on making discovery, as you know, for the last year, more transparent and predictable. And last year, we shared a commitment to you. And since then, we shipped a lot of improvements, providing essentially the signals that drive organic impressions. We've also introduced new tools to help you optimize those signals. For example, thumbnail personalization is used by about half of you now. And for those of you that use it, we're seeing about a 15% increase in engagement from recommendations for those that are using it. But wait, there's more.

Today, we're going to introduce a whole new way for people to discover and users to find your amazing content. It's Roblox Moments. It's a new app we're rolling out. Here's how it works. Users can capture great video throughout Roblox. With Moments, you can publish these moments. Moments is built on a whole new set of APIs using a feed system that will be really available to all of you. And something really crucial to think about here, we built Moments the hard way. We built it only with APIs and cloud tools that our developer community will have access to as well.

I want to invite our VP of Product, Raj Bhatia, to join me on stage, and we're going to do a little Moments demo. Come on out, Raj. And I think we're open sourcing this as well. So anyone can build it.

Anyone can be able to see this.

Okay. So let's dive in. Let's take a look.

Yes. So you guys ready for demo? We're going to do this live on here. All right. Cool. So I'm really excited about Moments because it's all about creator -- community-driven discovery. And it all starts with UGC captured game play video. So I think the first thing to do is actually to jump into an experience.

Let's go.

And let's capture. All right. Cool.

I want you get native capture inside any Roblox experience.

Yes. All right. So I'm going to hop into this island happen on this Island Obby. The team has challenged me to try to perform an Obby.

Okay. You're going to get to see Raj doing on Obby live. Let's see how good Raj is.

All right. Cool. So before we jump in, I think people have probably seen this, but you can -- we got a gallery for captures and then you can enable captures here through the hamburger menu, right? So I just turned on captures. You can see on the right side. The bottom icon here, take a screenshoot and I just took that. But we also have 30-second video capture happening as well. So I going to go and hit go.

Okay. Let's see what you got, Raj. Okay. Already better than me. Good. Nice.

All right. So we're capturing live. I can stop this any time I want. And of course, I fell in the water, right? So yes, cracking under pressure. So we'll stop that here, and the capture is now saved. If I want to see that capture, I can head up into the Roblox menu and hit captures here and you can see that my capture is there. I can play it back, I can delete it, et cetera. Why don't we hop over to Roblox Moments?

Yes, let's check it out.

All right. So I'm going to leave this experience. And because Roblox Moments is an experience, I'll search for it here. And all right, here we go.

So 100% built on APIs that will be available to creators.

Yes, we're releasing them...

Including the feed system.

Including the feed system, yes.

Okay. And what are you doing right now?

Well, I just loaded up Moments. I've got a moment playing here that someone else has uploaded. Let's create our own. So the first thing I do is I hit that plus button in the upper right-hand corner, and I'm going to allow Roblox Moments to access my captures and my media gallery. So I'm hitting allow. And you can see that, that same gallery is now available here. I'm going to pick this...

So once again any Roblox experience, we'll get permissions to ask to use your capture?

Yes, exactly. All right. So I want to use this. Let's hit continue at the bottom. I'm going to trim this video a little bit. Maybe I'll cut out that last part or maybe I should highlight it because it's how I feel, right? All right. I'll hit done here, pick some music.

Once again, any Roblox open source music here, I assume.

Yes, from our music library, yes. All right. Let's see. I'll use this one. All right, cool. Just trim normal. And let me add a description. Let's see, Kraken under the pressure.

Okay. So you just made an interactive video-based moment with music, trim the video, and you're pushing it now live.

I'm hitting post.

Okay. Cool.

And my moment is now publishing. It shouldn't take too long. It's going up to the cloud. It's video transcoding.

These are premade moments now that we've already made.

Yes, it's going through safety moderation.

Can you just maybe show the perf of that and how fast this works on your...

Yes, it's seemless. It's amassing.

So beautiful.

Really excited about this. And then, of course, the big point of moments is to bring people back into experiences. And so you see this join button at the bottom.

Right. So unlike a typical video feed, what's special here?

First, we react, and then let's join.

Any experience, that's the moment you can join.

Yes. And we're attempting to join now and it's on it's way.

Okay. That's awesome.

All right. That's moment. Thank you.

Raj, because we're trying to do everything like in the next few days, could you just push Moments live at noon today?

Do you think we're ready to launch?

Yes, yes, just go.

Okay. We'll make it happen.

Cool. Thank you. Okay. So we're hoping more users on the platform can find your great stuff as they use Moments to jump in. And we're hoping you can more and more start embedding video moments in your own experiences as well.

Now over the last year, there's been an explosion of new experiences on Roblox based on popular IP. Historically, IP licensing has been slow and complex. That's why we launched our IP platform. We started with 7 IPs, great partners like Netflix, Sega, Lionsgate. And our vision has really been to connect thousands of creators with thousands of IPs automating the deal-making process.

Now there's a specific type of content. It's really popular on Roblox, Manga and Anime. It's a global phenomenon. The revenues are expected to exceed over $100 billion by 2030 in the genre. And in Japan, Kodansha is one of really the big 3 Manga publishers, literally inspiring millions of people with their content. Today, we're announcing a partnership with Kodansha and highlighting 2 of their most popular IPs available for Roblox creators to apply today.

First, the time I got reincarnated as a slime and second blue lock we're inviting a very special guest to join us. Please welcome the CEO of Kodansha, Yoshinobu Noma.

[Foreign Language]. It's so great to have you here. Thank you for coming.

Thank you very much. I'm very happy here today and my first of audition.

Thank you. So -- Kodansha has been a leader in storytelling for over a century, and we have deep respect for this beloved IP, with our IP platform and companies like Kodansha, we're hoping to open a new world for UGC content.

I'm going to say this in Japanese, [Foreign Language] At Kodansha, we are embracing the future of fandoms and license 3.0 on Roblox. This model empowers the UGC community to create the final content under our standard, a new model our authors and industry partner agrees can coexist with the traditional gaming licensing.

At Kodansha, we are embracing the future of fandoms and license 3.0 on Roblox. This model empowers the UGC community to create the final content under our standard and new model our authors and industry partner agrees can coexist with the traditional gaming licensing.

Thank you so much. So this is a shared vision for both of us, empowering our creator community while respecting great IP, what do you think creators have to look forward to?

[Foreign Language].

We are treated by the enthusiasm the community has shown for our properties. We are only getting the study with 2 of our major franchises, Blue Lock and TenSura, available for application Roblox license manager. We are incredibly excited to bring in many more titles when Kodansha to Roblox in the near future.

Noma-san, we are thrilled. We thank you for your vision and your partnership. Thank you.

Thank you so much.

So we're really in the early phases here of IP on Roblox. We have a -- there's a pipeline of IP coming to the platform. Mattel is going to be bringing some of their 4 great IPs, Polly Pockets, Street Sharks, Rock 'Em Sock 'Em, and something I played a lot with when I was young Matchbox and Lionsgate's bringing Strangers, Blair Witch and Fall and many others are coming. Just thank you for all of our new IP platforms. Great to have you.

Okay. So we have a big show for you today. We shared a vision for safety and civility, not just on roadblocks, but really our vision of creating an industry best practice. We highlighted our technology and AI creation vision. We've introduced moments. We've given a glimpse of IP and content on Roblox. What we forgot the economy, so let's dive into this for a sec.

We shared our goal really when we talk to Wall Street and everyone to run roadblocks as efficiently as possible. So we have more money to distribute to the creator community and literally inside the company, we do watch every penny. So for the first time in 8 years, we're increasing the DevEx rates. So starting today, starting today, creators will earn 8.5% more DevEx on average for every Robux they earn. This will mean an additional literally would have meant 95 million more in DevEx over the last 12 months ending in June of this year. Thank you so much for your creativity.

We have a great big day. We have a stability and safety live AMA tomorrow, of course, open mic with yours truly. And I'll be back at the end of this morning with our top 10 predictions I hope you all have a wonderful morning and a great RDC. Thank you.

Please welcome Vice President of Product, Rajiv Bhatia.

Okay. Hello, RDC. It's good to see you all again. I feel like it's been a while since been up here. I'm Raj and I work on our Roadblock app, social and discovery products. And I'll be honest, I'm just really glad I got through that demo. So hopefully, I'll enjoy that.

When I'm not building demos, I spent a lot of time with the team thinking about how to empower all of you to reach your audiences in new and novel ways. And I think we've got something really special over Roblox moments.

So organic community-driven discovery has always been a key activity amongst our creators and players. We know that the Roblox community loves to share what they're playing and share the amazing gameplay and social interactions that you all have created for them. We also know that our community gets inspired to find and try new things based on what the community is talking about.

And they do this a lot. It's already on the screen, but in the short time that captures have been available, users have taken over 930 million screen shots and over 240 million videos within roadblocks experiences. This level of engagement shows our community is eager to share their best moments. And we think this is even more apparent when you consider that there have been over 1 trillion Roblox related views on YouTube. Just take that in for a second. That's a ton of views and quite an accomplishment.

So what you saw Dave and I just demo earlier is moments, a new Roblox experience built around user generated and shared content. And yes, as Dave requested, we're rolling out the beta today for 13-plus users at noon. So get ready for that. We think it's going to be a powerful engine for content creation and discovery, driven by your players. It enables them to capture, edit, share and discover the best in experience moments, the game winning plays, the Epic victories and the hilarious fails. Moments that others will find and enjoy, moments ultimately will bring them back to the engaging experiences that they originated from.

Now in true Roblox fashion, this is all built using platform APIs that you'll soon have access to. We'll also be sharing the moments experienced source code, so you'll all be able to integrate any part of what we built into your creations directly. So let's explore some of the possibilities that open up when you leverage these new APIs. We built some fun examples on the side just to demonstrate the possibilities.

Okay. First, let's look at how you can highlight and celebrate player achievement in your experience. Think about Avi players are completely focused on achieving a difficult sequence to clearer level. I know I was just trying that on stage. We're not trying to think about how to capture a highlight. We're thinking about making that next impossible jump, right?

So this is where the captures API comes in. You can automatically record those peak moments of gameplay in the background and then use the upload API to ask your users if they'd like to share that clip out with other people and without -- it never will break their focus. And then the bragging doesn't stop there. You can then use the recommendations API to populate an in-game highlight reel that you can show between levels inspiring players with the best performances of the day, they'll want to stick around to beat.

Now let's take this a step further and use these APIs to create something even more integrated into your game. While the previous example implemented the highlight reel in the games UI, let's make the communities' best moments as part of your 3D world. So imagine the dynamic billboard on your racetrack that uses a recommendation to API to show players the coolest overtakes or the fastest driving lines from other players. It's not just a static leaderboard for one player to see. It's a living highlight real embedded into your world, providing real-time gameplay inspiration for other people sharing the space together.

I know I love watching fast cars, just check out these drivers and -- before they start their laps. And then finally, we believe Moments isn't complete without giving players the ability to creatively edit their captures. This next example explores an even deeper integration of moments, one that becomes a core part of the GameLoop itself. In this fashion experience, the player isn't just capturing a moment, they're making it their own. Their striking a pose, they're taking their Moment, and then they're creatively editing and adding their own flare and customization before they publish it out for everyone else to see. We've included a few foundational editing features in Roblox Moments. And given you'll have the source code, we hope that you'll all use those or extend them.

I'm looking forward to seeing what new off-wall editing features also come from this community Okay. We're just scratching the surface of what's possible with these examples. We're showcasing short-form video in our Moments experience, but 3D video, live 3D. These are all formats that we're thinking about for the future.

And I want to remind everyone, it may not be apparent with these examples, but these platform capabilities are built and set up to work with any creation on Roblox. So even if you have an asset on Roblox, you can use the moment system with it. So I encourage you all to check out the Moments experience. It's going to be available starting at noon today. You can search for it or you can see it in the MR tab. And if you haven't already, get your players capturing and sharing from your experience by integrating the video captures API. We can't wait to see how Moments will boost creation, social interaction and discovery across the platform.

I'm looking forward to seeing how all of you innovate with this, and I encourage you all to push the boundaries. Okay. So shifting gears, we just covered how getting users to share their best modes can be a powerful way for you to engage and reach a wider audience. But that's just one part of the equation. We also know that to grow in the next decade and to reach 10% of the gaming market, we need to bring Roblox and your incredible experiences to more devices, meeting players where they are.

And that's why our vision is for you to reach every player on any device anywhere. And with the platform expansion investments we've made over the past few years, Roblox is now available in more places than ever before. The growth across desktop, mobile and consoles is a powerful testament to your hard work and the high-quality experiences that you're building. Your creations are driving more engagement and connections between players across these platforms.

And we've talked about this in the past, but to remind ourselves, we've seen that users that play across platforms engage 20% more and spend 2.5x more on average per hour than those who play on a single one. We know the promise of game development on Roblox is big. You build a game, you click publish, you can be on all these platforms and you can reap these benefits of cross-platform play. But we also know that this can be a challenge. Getting your creations to be amazing across all these devices does take work.

And this device optimization journey is key to making your creation seamlessly accessible by that ever-expanding audience across the ever-expanding set of devices that we're on. So we're working on making this easier for you. Just a few weeks ago, we shipped a toolkit to help you in this journey. Our adaptive design principles give you the guidance on what to do. New and updated APIs provide practical tools to do it, addressing key elements like display, input and styling across device types. In our studio emulators, you'll see that we've added a new one, enable you to test your game specifically for new form factors.

This toolkit is designed to help you build an optimized immersive experience that truly sings on any device. And we're not stopping here. On October 16, -- the new Xbox Ally handhelds are launching, and Roblox will be playable natively on day 1. This is about delivering true portability and performance, putting Roblox in players' hands wherever they go. I'll call out one other interesting thing about this device. It's going to present an interesting challenge for us as well.

This is a portable device. It's got game controllers. It's got a touchscreen. You can plug it in and dock it. The tools that you need to be able to make your experience thing on all these modes, we have that toolkit for you. So we can't wait to see this community's creativity shine with the power and fidelity of this dedicated gaming handheld. Okay. That's it for me today. But if you're ready to dive deeper on anything we covered, we've got 2 great breakout sessions tomorrow. learn more about reaching every player and optimizing for all devices and the handheld at 12:15 p.m. or learn about moments and building connected experiences of Roblox APIs at 1:00 p.m. Thanks, everybody. Have a great RDC.

Please welcome Roblox Senior Vice President of Engineering, Anupam Singh.

Another great year at RDC. How is everybody feeling? Last year, we talked about 3D and 4D generative AI. And you saw Dave talk about this in his portion, where we believe that player-generated content using generative AI is going to be a fundamental game changer for us as a platform and for you as creators. Now I know you did rather hear about this from a creator. So I have a question for you. How many of you have heard about Jailbreak I would like to invite Alex who's created Jailbreak is already successful on the platform to tell us more about generative AI. Alex come on over.

All right. Alex, let's start with -- let's start with seeing how you've integrated the Cube AI technology into Jailbreak. Let's queue the video.

[Presentation]

So Alex, how does custom content generated by players change jailbreak?

Yes, I was surprised. I had seen the technology being developed, and I played with some image generators, but I honestly didn't anticipate it being so fun myself to be able to take an idea in my head and so quickly, be able to actually drive it. And it kind of gives me the feeling that I had 15 years ago when I started making games.

And let me play the AI skeptic for a while, right? Player-generated content is very well understood. What's the difference between using the Cube or not using the Cube for players?

Yes. So prior to this technology, there was no way for most of our players to get their own ideas and their own car mesh into the game. For the past few years, every 2 months, we've held a Jailbreak season contest, Jailbreak creation season contest, where we allow users to submit their ideas, and it's hundreds of users that submit their ideas, and we choose 10 every 2 months, and we put them into the game. And we're amazed that so many passionate players have taken the time to learn how to 3D model in Blender, in our style, and we'll continue to do this but we're very excited because we know that, that's still just a fraction of the users that we have playing Jailbreak that would love to see their ideas brought to life in Jailbreak immediately.

Okay. So 10 pieces of content with a lot of effort are trying to keep the style consistent with Jailbreak compared to what we just saw. Now we are all builders here. We're all engineers. How did you integrate the Cube APIs into Jailbreak?

Yes. So we utilize the multi-mesh geometry generation. So we're generating the car body and the wheels separately. And we're using the positions of the wheels and the wheel mesh itself. So we can still have the wheel spin on the ground, and we can still have the suspension work, and it's awesome.

Thank you very much, Alex. This was -- hopefully, everybody get to know something new. Thank you.

Thank you.

Now we all know when AI-generated content becomes easier to generate a lot of us as creators and engineers think about how do we control this? How do we know that it will match the style, the behavior, the size, the aspect ratio, how will the experience remain consistent? To find answers to these questions, I want to invite Kiran Bhat who is the lead engineer for Cube API. Kiran, come on over.

All right, Kiran, what are we doing today?

So I hear from you live demos.

Yes, yes.

How about a live game instead?

Oh, my 18-year-old thinks, I'm a really bad Roblox player. So let's try them.

Yes, let's jump right into it. So what we are going to show you here is a capture the flag type game, where one of our team members built all the game assets using the Cube AI. And as the creator of this game, we wanted the players to build all the objects for the game play like vehicles, planes and even guns that can shoot using the Cube. So let me start with a powerful vehicle that can drive, so maybe a Battle Mech. So let's try a Battle Mech with if I can properly for wheels. So I'm going to just generate that.

And one thing is, as a creator, you get to specify the overall style of the game and the objects. So let's see, if I switch to this Sci-Fi robots theme, the same prompt gets interpreted by the cube into this future industrial Battle Mech, which is more Sci-Fi theme. So I can generate that. And while we are at it, there's also another theme here, which is cute robots animals.

So why don't we give that a try as well. So now the Battle Mech becomes a bright yellow hamster. So let's generate that as well. So we have these generations going. So what's happening here, if I'm getting it right, is the creator, in this case, maybe it's you, has set up a particular context in which we will generate the AI object. That's right. And those context can be changed and CBI is smart enough to apply that style rather than generate some jump.

That's exactly right.

But are there any other controls for creators?

Yes, that's right. So one other big control is the ability to specify the size and the aspect ratios of the objects that are getting created. So let's try that. So if I maybe stick back to the Sci-fi robots. And if I add a bounding box here. So now I'm going to try and change the base of the Battle Mech and make it a little taller. So I can go ahead and change this box a little bit here and then maybe make it a little taller. So I do generate again. So it goes back to things and then the generation is going to fit into this box.

But resizing objects is nothing new on our platform, right? You just resize something. Why is this important? This is actually not resizing. So what's happening here is that these bounding box dimensions are fed as inputs to the cube model along with the prompts. So the output as it generates is going to adhere to the prompts and try to fit within the bounding box. So it's essentially a form of control where each generation is going to be different depending on the size of the bounding box. So Anupam, what did you land up generating?

I generated a short car, where the car has wheels, but it is also a shark with bare teeth. So it's pretty fine to show it to you guys. That's the shark [indiscernible] with bare feet. A round of applause for our Cube AI.

Right this car right there. Okay. So one interesting thing is, okay, here, we see all our Battle Mech and let me turn this debug option on. See how -- this is our cute animal teamed Battle Mech. This is the futuristic Sci-Fi team Battle Mech. And let's go and drive one of these guys. So this is one key change that we have introduced. The objects now are what we call 4D, which is 3D with embedded scripts, which allow them to be drivable. So if you create buildings, you can maybe have doors that open. And here, let's go and drive this thing here.

As you're driving it tell us more about how did Cube AI understand that it is a drivable thing.

Yes. So it's -- let me just show you that. And also one other thing is we can -- it shoots, right? So it's both drivable and shootable. So let's look at the reasoning that goes on here. So the prompt was the bottleneck with 4 wheels. And the cube reasoned about the prompt and decided that it has to be both drivable and shootable, right? And this is a really important thing. The basic capabilities, whether it's drivable or shootable or flyable is set by you as the developer. And the model is just taking -- interpreting the user's prompt, the players' prompt and then composing the behavior for the generation. So here, it made it drivable and shootable. Let's take a look at your creation.

My shark can shoot?

Let me see. So this is your shot here, right? So it says it's just drivable. So even though it is shark, it decided that the behavior is not shootable.

So one of the things for creators in the room is that you can actually see the reasoning used by the AI to build the object. So not only does it have behavior, but in some ways, you can control the behavior and so you can control the experience for all the players while the players can continue to generate AI-based optics.

So we also have a bunch of guns here. It's a large gun that shoots watermelons.

This is the same watermelon gun that Dave and Kiran right?

So I can shoot watermelons here and see how our friends from the other side are coming. So let's...

Oh, you're trying to shoot.

All right. There you go. That's another shot to shoot. Anyway, I can play this all day long.

We could do this all day long, but I'm getting the queue that we should get our stage. So what is the takeaway? The takeaway is creator control is the most important thing for us in AI. We want to build everything such that you as the creator can control the style, you can control the size and you can control the behavior so that the AI-generated objects are still objects that you would want in your experience. With that, thank you very much, Kiran. We played Roblox.

[Presentation]

Please welcome Roblox Vice President of Global Brand Partnership, Stephanie Latham and Vice President of Product Ads Louqman Parampath.

Wow, when creators and brands come together, magic happens. Hi, everyone. I'm Stephanie Latham, I lead our global brand partnership team, and I'm so happy to be back together with all of you again this year.

Hi, everyone, and I am Louqman Parampath, the VP of Ad and Brands product at Roblox. This is my first RDC and I could not be more excited to share what we have built with all of you today.

Me too, we love working with you to build products that help you both succeed and grow. And as most of you in this room know, we intentionally design all of our advertising solutions to be a win, win, win: a win for creators, growing for brands and, of course, a win for our users.

Creators can grow their audience and earn revenue to help take their experience to the next level. For brands, they can connect with Roblox as engaged community and build new fans. And users get entertaining content and useful in-game benefits, making their experience even greater and more engaging.

We're investing in ads because we want to help you all unlock the full potential on our platform and beyond. So let me show you how this works. First, ads can help you get discovered and grow. Running ads can bring new players into your game and keep them coming back. In fact, did you know that the 5 newest experiences to break into our top 20, all started with the help of ads.

But that's not all. Ads can also help you earn more. placing ads in your experiences can help you monetize more of your users. And we want to help you do this thoughtfully alongside subscriptions, in-game purchases and more. It's a cycle that builds on itself. More players, lead to more earnings, lead to more for you to reinvest. And here's the cool part, once you reach scale, then you can expand unlocking commerce, licensing opportunities and, of course, direct partnerships.

We've invested a lot to make this lower reality this year. Haven't we, Louqman?

Yes, Stephanie, we have. Let's start with how we help your creators grow on our platform. This year, we upgraded Ad Manager, your tool for running ad campaigns that help you reach and retain more users. We made Ad Manager more intuitive. Just tell us your goal and our AI will help find players aligned with that goal. We shipped over 40 new improvements to help you get more players at lower cost in less time than ever before.

Campaign creation is now much easier and can be done in just 3 steps. Just put in your goal, your budget and your creators and you're ready to go. The cost to get a player to join your game or cost per play is now much lower, 71% lower, thanks to performance improvements under the hood that we have made. This combination gives you more time and money to invest back in building better and greater games.

And we love the adoption that we are seeing from the community already this year. Over 45,000 of you have used Ad Manager in 2025 to grow your games. Whether you're finding new players, keeping current ones or bringing back old ones, we are thrilled to be a part of your success. But we're just getting started. There is a lot more to come.

There sure is. Next year, our new maximized spend objective will automatically find the players, most likely to spend your game. And soon, we're combining creator analytics with as reporting. So you can see the true impact of ads on your entire experience. And creators, we heard you too. And in the next few months, we're bringing the power of the new ads manager to our search ads. To ensure that your game is promoted right when a user is looking for something new.

Now that you grow your game, let's talk about how we can diversify the ways you earn from your game. We are building for a world where you can make meaningful -- ads within your -- with full control over how ads sit alongside or is additive to your other monetization tactics.

And it's not just the ways you own, it's also who you earn it from, whether it's new users, casuals or top fans of your game, we want to give you the tools that enable you to build the right choices for all users into your game so that you can maximize your earnings. For example, this year, we launched on much format, rewarded video ads.

Over 100 experiences are already using them and the results speak for themselves. The creators of Blue Haven are already earning 5 figures monthly within 2 months of starting the program. And as we saw in the video before, creators of other games such as EASY Glass Bridge have seen earnings showed up by more than 40%. So today, we are really excited to share that rewarded video ads will soon be live for tens of thousands of more creators to use as an additional way to earn on Roblox.

Now we have also invested in giving you more controls, more functionality and more reporting so you can better manage your ads and earnings. This will make it easier for you to find what works with your ad campaigns for your game and your community so that you can continue to expand your game.

So many great innovations, but we know and we love that this crew isn't just making great games. You're building and creating the future of entertainment. And as Dave mentioned earlier, when you pair up with brands, real magic can happen, and this year continue to prove it. We enabled commerce our API. So creators can now sell their own products on robots. Love your hoodie, Louqman.

Thank you. Shoutout to the creator scenario that I got it from.

And we brought this to brands, too. Fenty launched an exclusive Lip gloss on Roblox, and we didn't stop there. We had Travis Kelce running Roots and Girl Garden. And we brought Jurassic World and their dinosaurs to the homepage of Roblox. And just the other weekend, we had Lady Gaga bring Mayhem to the platform. Culture is happening on Roblox, and you're the ones making it. So we're excited to keep building the tools that help you create this win, win, win and keep building the future with you. We hope you'll join us tomorrow to learn more at the Science of game discovery and mastering monetization breakout sessions. Thank you.

Thank you, everybody. Please welcome Chief Marketing Officer and Head of Market Expansion, Jerret West.

How's everybody doing this morning? Is doing this morning? Give me a little bit of juice. I need to level up here. It's amazing to see you all. Thanks for being here. And I love this conference. I love walking around like at the happy hour last night. And I was thinking about why that is, why this place is so special. And it's because when I look out at everyone out here, I see an idea. I see an idea that you believed in so passionately that you decided to turn it into a creation. And you decided to put that creation on Roblox for which we are incredibly grateful.

That's incredibly rare in the entertainment space to have that kind of passion fuel a platform. And then I think about a little bit further and you made an incredibly smart financial decision as well. Roblox is platform. This platform has been growing at about 8x faster than the video game industry. It's been that way for quite some time. So I ask myself, how is it? Why is that the case? And I keep coming back to one specific thing. Out of all the platforms that are out there in the entertainment space, I just don't think there's a single one that is as closely connected to the creator community as the Roblox platform is.

Our growth is connected. We scale together. The platform supports your creations and ideas and your creations make Roblox better for everybody that's playing out there. It's a wonderful virtuous cycle. So I'm going to spend the next couple of minutes talking purely about things where we can enjoy mutual success, where we can grow together and share some ideas on how we could do that. Roblox initially started as a U.S.-centric product, obviously. It's grown to the exact opposite. 80% of our daily active users are playing from outside of the United States at this point. So it's incredibly important for us to figure out how we break down those barriers together, one of which is language.

And our vision here is for language translation, automatic translation to occur in real time across voice, text and image. In order for us to get there, we need you to participate, and we're off to a really good start. About 85% of our top experiences today are using automatic translation. This is one of those things. When you're creating something new, hit the toggle in the Creator hub and turn that on. If you've got a long-tail game that's not sitting in that top 100, hit the toggle in the creator hub and turn those games on because that long tail helps us get there faster.

With the way that we are scaling and growing right now around the world, automatic translation and that notion of seamless communication, no matter what language you are speaking is more important than ever. So let's start there with automatic translation. The other part of this is making affordability accessible to everyone everywhere. And again, this is rooted in our origins of being a U.S.-centric product. Oftentimes, that would establish the global pricing. That reference pricing isn't great for a lot of places around the world.

So it's imperative for us to be able to come up with regional pricing. We've got an option here to do that. And it automatically in real time gives you pricing preferences or suggestions that you can take or leave for anywhere around the world. All you have to do is turn this on the Creator hub as well. And the vision here is to make great gameplay accessible for everyone around the world no matter where you are or where you're playing from.

Okay. So the third piece here is just talking about the 138,000 servers that are online and ready to roll right now. They have specifically been built for moments like August 23 when 45 million players showed up at the exact same time to play a couple of games. Shout out to the infrastructure team for standing that up. I'm not sure that would have been the case 5 years ago, but it does just kind of indicate the astounding progress that we've made on the infrastructure side.

Pretty incredible. Now is the point where we get to give them a heart attack because the call to action here is for you to do more of those, bigger. We need them. We need them to fuel our conversations. And there's one big reason why I think about this as a marketer. Those are the moments that lock in core memories for all of our players around the world. Those are the things that they talk about. Those are the things that become things that are talked about across the Internet.

And we need to turn that into cultural fuel. In turn, -- we also need to have the infrastructure to be able to amplify all of those ideas and creations for you. That's why we've spent the better part of the year focusing on building out the Roblox marketing engine. And simply put, this is really all of the infrastructure or areas where we can do our best work to amplify all of your stories and turn your creations into those cultural moments. We started the basics with social media. So we focused a lot on increasing engagement and increasing our follower count around the world.

We're up to 48 million followers on our handles. That drives about 70 million impressions and 4 million engagements a month for you all, specifically focused on boosting your content, boosting your game handles or bringing new players into the Roblox platform. So making tremendous progress here. We are road mapped out to get to 100 million followers. We need your help in doing that. We need that crosstalk across your handles and you sharing with us what content is going to be big so we can turn them into great stories on social media. Perhaps the inverse of this is thinking about these long-form stories.

There are many times where we need to talk about elements of the Roblox platform and ideas shaping that platform. But more importantly, we've got moments where we want to share your creator stories. It's amazing to talk to you all one-on-one. We want to take those opportunities and turn them into podcast moments where we tell a little bit more of a creator story and we tell a little bit more of a content story. Today, we've got 3 million listeners already on our Roblox podcast network. We're going to be adding a whole bunch of new formats, and we're going to be reaching out to you all to help us tell those stories in the coming months.

But more importantly, we've got moments where we want to share your creator stories. It's amazing to talk to you all one-on-one. We want to take those opportunities and turn them into podcast moments where we tell a little bit more of a creator story and we tell a little bit more of a content story. Today, we've got 3 million listeners already on our Roblox podcast network. We're going to be adding a whole bunch of new formats, and we're going to be reaching out to you all to help us tell those stories in the coming months.

All right. Another part of this engine has been about building the top of the funnel. We are up to about 420,000 retail locations in 45 countries with Roblox [ pegs ]. It's been a tremendous amount of work to build out that experience and presence. And in addition to that, we've also built out a relationship with all of our digital platforms, store platforms around the world. That's reaching about 2 billion impressions a month as well. So the next couple of months is going to be about building the clock speed of turning and chasing heat around the content that you all are generating.

So look out for us to be able to pull those ideas and experiences into our retail and also into our digital store platforms. Another area that I want to tell you about, and this is one that you can literally walk out those doors if you're interested and visit the advertising creators' booth. We have a strategy around paid ads. You just heard a lot about our advertising technology and branded advertisers technology. That has created for us an amazing targeting technology, an amazing ad tech stack and it's great, great data store around how people are behaving on Roblox. What we want to do is enable that so you can build out paid media programs and media plans that run targeted external ads for your games. So we'd like to drive on and off-platform new users to your experiences using the same advertising technology.

Like I said, if you're interested in that, go hit the advertising for creators booth, they'll be able to tell you all about it. All right. Last part of the engine here is probably the most important. And I want to recognize a lot of the people whose names or handles are on the slide here. They're in the room with us today, but these are our influencers, and we're lucky to have an incredible community of influencers. Some of you have been with us for years, and your creativity continues to shape Roblox. So thanks for the content you're creating, the energy you bring and in many instances, the feedback that you share. You've been a vital part of Roblox success. And I hope that influencer and creator have an opportunity to meet increasingly at events like this.

Just to underscore how much reach and cultural relevance this drives for us. And again, I'll emphasize the fact that most of these videos that we're talking about, the video views here are really talking about your experiences and content. We crossed over the 1 trillion video view threshold on YouTube. You can count on 2 hands the number of brands that have reached that threshold. And we're accelerating about 1.2 billion video views a day on YouTube and TikTok currently.

Talking with our partners over at TikTok, they've told us that about 20% share of voice, 1 in 5 videos on TikTok this summer have been Roblox conversations, mostly about your experiences and content. So exciting stuff. We've got an apparatus here to amplify your stories, and that's really what we have built this for. And we amplify the engine by building really fun fuel for it. And I'm going to share a couple of things that we've observed that drive big conversation that you all can hopefully draft off of on your handles or you can partner with us to build out even bigger. One is nostalgia. Roblox is 19 years old. We've got a big one coming up 20 years next year, but 19 years old and has nearly 2 decades of memories inside jokes. I mean it's no longer a platform. It's a part of our lived experience, part of our story. I'll use one example for this, and that's when we brought back the oof sound. It's almost primal for a lot of our users, our players.

95 million views on Instagram and TikTok for oof. Same thing for Cheezburger. We had to cut one. I went with oof. Meme generation is another area where you can explore and think about a little bit further. Don't be afraid here. It has to be authentic. It has to be super timely. But if you pull it off, bringing in real life or digital memes into the Xbox world and vice versa, super powerful. We're an entertainment platform. This is where you have a little bit more license than a traditional brand or a traditional experience. to play with. Take, for example, Aura Farming Kid. Yes, I'm not sure why he's a Canadian Rockies Lumberjack in that video, but 22 million views on TikTok and Instagram. So trends could start anywhere, but you can ride them with creative here and timely updates.

The most important piece here, though, is your creator updates. Can't emphasize this enough, have a point with your creator updates, have a pitch angle, have a hook, have something that brings it back to life. Work with us so we can amplify those things. And you'll get real magic to continue to extend the life. Example being Style Showdown. That was March. This beta dropped, fresh format, daily play, good hook angle. If you ship a great update, we'll help turn it into a cultural catalyst like this. And we want to create more of these moments that we can flywheel and keep the life cycle of your experience going beyond just launch month.

All right. Let's flip the page and talk a little bit more about how we are leveraging huge events, how we are creating them because we're going to be expanding our playbook in this space as well. We've got some classics here with our live ops events, right? So far this year, we've run 2 events. One was The Hunt: Mega Edition, which was wildly successful, super fun, right? And The Hatch, that happened. And I'll tell you what, though, collectively across the both, they both achieved what they were meant to do. 346 million hub views across the two, 87 million players participating, doing a great job of engaging, helping people with discovery and drive people to your experiences even if they're under recognized out of the game.

But I've got something here for you that I want to tell you about, and that's our third event. It's a brand-new concept called The Takeover. And it is a coopetition in a brand-new place called Tagtown, you're going to have the ability to bring and build crews and you defend the neighborhood against other crews that are built and doing everything that they can to build a bigger network of neighborhoods, if you will. This is an incredibly unique art concept, as you can see. It launches on September 12, so mark your calendars. I'm really excited about this one because I think it's going to add to our Live-Ops roster. This is going to be a really, really fun one.

All right. So like I mentioned, it's coopetition. This is one that brings people together around that co-op gameplay and competitive spirit. We're also expanding our roster around things that merge the joy of immersive play on the Roblox platform. Of course, I'm talking about music here and concerts. Sophie Tucker, LEV, Kpop Demon Hunters all have been on our platform within the past months. And now we're leveling up and creating an experience called The Block. It's our new persistent event space, and we're partnering with creators across the community to bring these exciting events to life. So we've got one coming up on Saturday tomorrow. BSlick is going to be performing in real life at the RIA's after party, and then he's going to be in The Block performing for our Roblox users that aren't going to be like us at the RIA's.

In addition to that, next week, we have aespa coming to The Block. Yes, they're going to be showcasing their virtual item collaborations with 4 creators. And this is really just the beginning. We're going to be bringing more artists and more events to The Block in the coming months. So really excited about that. Lastly here, we can't forget about the final for The Hunt: Mega Edition. And that really lit the way for something that we think we can take advantage of in 2026 and lean into, and that's esports and competitive gameplay. As you saw, our engine is growing by leaps and bounds. It looks incredible. And we're going to lean into esports. So stay tuned for this, start thinking about leaderboards, tournaments and game modes right now for your games because we're going to be looking a little bit harder at this.

All right. So that's my time. And I want to say we are just -- we want to be your biggest champion. We're here to spotlight your creations and help them shine across the world. And if I say one more thing about RDC, I just want to encourage you all to do everything you can to step back over the next 48 hours and look around. You are building -- you are building with friends, sometimes family, you are building something that you're passionate about. We're in a space that's growing exponentially. It makes me think of that classic quote from that old sage, Andy Bernard in The Office saying, "I wish I could look and have the ability to see the good old days while I'm living them because that's really what you all are in right now, and I hope you take advantage of it. I look forward to meeting you over the next couple of days and have a fantastic RDC.

Please welcome back David Baszucki.

Okay. All right. It's almost lunch time, a couple of caveats. This afternoon is even more technical, more engineering, more product, and more scale than you saw this morning. So stay tuned. It's -- we have a huge afternoon lined up for you. For those of you at your first RDC, tradition time, every year, we've been doing top 10 predictions. We've already done one cycle. We're going to do another cycle. So let's do what we do every year. Let's bring up the predictions we made last year, and I'll give you an update.

First prediction. Fortune 500 company will use a Roblox experience as part of a recruiting process. Well, I will highlight that we're using Roblox as part of our recruiting process. Now I know there's a lot of investors in the audience, so we don't want to make any forward-looking statements, but I'm very optimistic about this one, and it might be Roblox. Next up, a school will integrate a full K-12 curriculum. For those of you that are watching, we introduced a full educational discovery sort. People making educational experiences are getting more and more visibility. I'm still super optimistic about this one.

The next one, let's up this to 18 plus. Okay. So this one is controversial, but here's what I want to share with you. We've seen mail system, we've seen text system. We've seen voice system, we've seen video system. 3D immersion is inevitable. It's going to be a communication platform. And I'll flip this in reverse. We see this as a corporate responsibility to usher this technology in with safety and civility. So with 18 plus ID verified people on our platform, we are optimistic about this.

Next one. Roblox employees, more time using Roblox for meetings than with video. It's early. I still think we can do this one, so I'm going to keep a thumbs up on that one. Next one. So I'm pretty optimistic this is already happening, and we're going to go and measure this. I talked to a lot of parents with the nudges, with our safety and civility systems. I actually had a Roblox employee come to me whose young child, I think, was saying, not bad stuff, just edgy stuff. And for some reason, they got a 1-day ban and it created a discussion. So very optimistic that when we measure this, we'll be increasing the civility of people. So I'm thumbs up on that one.

Number five, let's take a look. Already happening. We already have creators. So this one should actually be done. Let's go -- oh, updated slide. Cool. Roblox will be a frequent communication channel for my family. I'm hoping we'll see, okay, my kids are a little like they've seen a lot of Roblox over 20 years. So this one will hope on. So we're going to kind of put a little warning on that one.

Number three, so go out in the lobby. I think it's already clear. Roblox is going to be a place for fashion design. I'm very optimistic. I'm hoping with some of the things we've seen with fashion on Roblox, it's really only early, so thumbs up there. Number two, absolutely, we're talking to people right now, building large festivals. We all know we need 3D simulcast technology. And when we have that, we'll be able to go live out to millions. So -- and the concurrency numbers of this actually look pretty small now, right? 1 million, like that's small potatoes now. And then finally, for those of you that track M&A, for those of you that track the value of studios, for those of you that track the value of Roblox and what we're trying to do, I absolutely believe we're going to see a $1 billion studio by the time we finish this 5 years. So thank you all. Really appreciate it. We have a huge afternoon this afternoon and really appreciate your creativity. Thank you.

Please enjoy lunch. We'll see you back here for the creator keynote.

[Break]

Please welcome Roblox Senior Vice President of Engineering, Nick Tornow.

Hello, RDC. It's so great to be back here with all of you. My name is Nick Tornow, and I lead our creator and engine groups here at Roblox. And I tell you all year, every year, this is the event I look forward to the most. And I know what you're thinking. But no, it's not because of the [ swag ]. So the [ swag ] is very good. It's because the energy, the collaboration and the pure creativity in this room is higher than anywhere else in the planet. This is what is happening. And it's this which fuels everything we do at Roblox.

So welcome to the Roblox Developers Conference. This is our time to connect, to learn and to celebrate everything you all have accomplished over the last year. It's going to be an incredible couple of days. Now we have a lot of exciting announcements to share with you. We're going to geek out a little bit. But before we look forward, I want to take a moment to look back at everything we've accomplished together over the last 12 months because what a year it's been.

The last 12 months have been a year of explosive growth, and it's been driven by your creativity and your dedication. Let's consider our platform growth. Our global community is more engaged than ever before. And we recently hit a new milestone of 111.8 million daily active users across the platform, which is a growth rate of over 41% year-over-year. Now when numbers get this big, sometimes it's helpful to try to put them in perspective by comparing to something else. And I want to call out the most recent Super Bowl had 127 million viewers. So this past quarter, together, we averaged approximately a Super Bowl on Roblox every single day. Yes. It's incredible. And these users are diving deeper and deeper into your experiences than ever before. Now there's a lot of ways to look at this. One way to see that is in the last year, we've seen more than 390 billion visits to experiences. Now this is up 41% year-over-year. And if you do the math, this represents about 12,000 visits per second, which, as I understand it, is about the rate at which people all over the earth enjoy Coca-Cola beverages.

It's also incidentally approximately the rate at which Roblox creators ask for emissive map support from our engine. Now another way to look at this is in the last quarter, players spent a whopping 27.4 billion hours in your experiences. This is a 58% increase year-over-year. And it represents an average concurrent count of about 12 million users. Now if you recall earlier from David's opening keynote, this is about where we were at the beginning of the year. This is about the peak Roblox had ever achieved at the beginning of the year. We're actually about 11 million. So that -- what was our peak at the beginning of the year is now our average for the last quarter. So users are playing more. They're exploring more and they're staying longer.

Now all of this is due to your hard work. You've been so productive this last year keeping content fresh and keeping users engaged. We see this everywhere we look. But just one example is that every day, 26% of the top 1,000 experiences are updated or published. And you've been busy building on each other's work. We see this with the creator store. You've inserted more than 1.2 billion models into your experiences, and you've installed more than 500,000 plug-ins since the last RDC. And I want to give a big shout out to all of our plug-in creators out there for accelerating creation across the entire community. Thank you.

All right. Now as the platform expands, so do the earning opportunities for all creators. Earlier, David shared that creators over the last year have earned over $1 billion. Our creator ecosystem includes many hobbyists. And of the millions of creators monetizing on Roblox, over 29,000 are part of our DevEx program, with the median creator earning $1,440. But looking a little deeper at those numbers. In the last 12 months, the average revenue for a top 1,000 developer was $980,000, so nearly $1 million. For a top 100 developer, that average was $7 million. And for a top 10 developer, the average was $38.5 million.

Now these numbers sort of speak for themselves, but I want to call particular attention while we see growth across the board, and you can see the growth numbers up on the screen. We see a growth rate for the top 1,000 of 2.9x over the last 5 years, which is higher than the growth rate of 2.8x for the top 100 and higher than the growth rate of 2.2x for the top 10. This means we're becoming a broader and deeper platform, and this is a trend we plan to continue.

Building further on this, we also launched Creator Rewards earlier this summer. And it's early days, but we now have more creators earning for engagement than ever before. And here's just a couple of stats. We're seeing an increase of 171% number of creators earning with Creator Rewards relative to the prior engagement-based payout system. And on top of this, we're seeing of those creators who are earning more than 5 Roblox through the EBP program, 73% of them are earning more with the Creator Rewards program. So more creators are earning and all the creators are also earning more. So this broad-based growth is what we like to see. More and more people are able to build thriving businesses on this platform.

Now underlying the strength are stable experiences continuing to drive growth. And here's another amazing stat. There are over 160 experiences on Roblox that have reached the 1 billion visits milestone. Now in this past year, 50 new experiences have joined this exclusive club, which is a growth rate increase of more than 50% relative to the prior year. But here's another thing to be very excited about. Beyond this core strength of stable experiences, which we know and love, this is not a static platform. In fact, it's a platform that is increasingly dynamic, and it increasingly powers breakout hits.

If we look just at experiences created in the last year, we see more than 2,000 of them have already hit 5 million visits. We see more than 1,000 of them have already hit 10 million visits. And amazingly, we see that 60 of them have already hit 100 million visits. This is within 1 year of that experience ID being set. So this means there is not a fixed set of leaders on this platform. New entrants have explosive growth potential. There's more experiences bursting onto the scene, but also core strength in established content, which is really the best of both worlds.

Now each time a new hit experience comes on to the platform, Roblox grows. And as the platform grows, this helps everyone else. We know this because the average user on Roblox plays 21 experiences per month. So players new to the platform might join for one experience, but they stay around to play many other experiences, benefiting the community as a whole. You can think of Roblox like a big tent, where each breakthrough experience adds a new tent pole, raising the canopy and inviting more users and creators to come in. And we're just getting started. You heard this morning about our audacious goal to become 10% of the global gaming market.

Now to get there, new tent pole experiences will emerge. New genres will emerge. And new genres push the boundaries of technology and game design, redefining our platform and breakthrough this ripple across the ecosystem. So let's take a look at how things are going at the genre level. For this next section, I'd like to invite Greg Hartrell, our product lead for creator content to the stage. Welcome, Greg.

Thank you, Nick. All right. I think we'll start with talking about shooters as a genre. Shooters hours played are up more than 100% year-over-year. Top shooters like Rivals and Gunfight Arena deliver the joy of a precision shot to secure the win. But they've done this by building their own server side hit registration systems. They've optimized their games performance to perform well on a Roblox game engine and made these games playable on more devices supporting different input types to reach a wider audience.

Now I love these 2, but I feel like I've seen some new ones break through recently as well.

It's absolutely true. So it's worth calling out weird gun game. It subverted this genre in really unique ways, building a part-by-part crafting system where you can assemble almost any kind of weapon. It's a pretty compelling game loop. You build some -- any kind of weapon you want, you jump into a match to test it. And then you get to see if you're kind of brand of weird wins.

So I really wondered, Greg, and maybe you can help me out, are these weird guns actually effective?

Well, sometimes weirdness is a strategy. I made a gun, for example, when you shot that made a loud hawking sound every time you shot.

A honk. Yes. So how does that help?

It's actually not very helpful. It turns out that your enemies know where you are all the time. But when you do this, your opponents know that they're getting killed by somebody with a honking weapon, and that's a psychological victory.

I see, its very deep. I have many questions, but I think we need to move on.

Okay.

How are some of our more traditional genres today?

We've got some really impressive growth in simulation, for example, in the last year. The genre evolved with the explosive growth of Grow a Garden and Steal a Brainrot. They've smashed concurrency records in their own ways, but they've each done some interesting ways of reaching new heights, adding mechanics that encourage and reward compound social expression and executing on these epic live events such as the admin wars that we saw only a couple of weeks ago. And I have to confess, I don't know what Torrtuginni Dragonfrutini is, but it makes me hungry.

Don't eat the memes.

Really fruits in the name, Nick, it's edible. Okay. But if we're talking about food, we are talking about another genre. We're talking about survival. And this is where we've seen survival explode, and we saw the gripping and strategic challenges of 99 Nights in the Forest, pushing the genre forward. Very impressive. It innovates by seamlessly innovating with survival crafting, a little bit of tower defense, some open world exploration, and it features procedural map generation and AI path finding and behaviors for its NPCs, not to mention the struggle with a Demonic Deer with boundary issues.

Yes. I know all about deer struggles. We have one that comes into our yard and eats our flowers, then it goes to our window and just kind of stairs at us.

That sounds kind of terrifying, Nick. Maybe the deer is looking at its reflection.

No, I'm pretty sure it's the psychological victory thing you were just talking about.

We got to fear the deer, Nick, fear the dear.

All right. Let's shift our focus to our road map. Our partnership with you all is built on trust, and trust is built on transparency. Right now, we're going to do a live retrospective of our Creator Roadmap. We need to look at what we said we'd do and what we actually did. We're here today to hold ourselves to account. So we launched our Creator Roadmap in 2023, and we update it every quarter to give you visibility and build confidence that we're doing the right things for everyone in this community. And since our last RDC, we've shipped more than 102 items on our road map, but the details are an important part of showing what kind of job we're doing. So let's take a look. Here's our report card, wins and misses altogether.

While it's great to get 102 items out to the community, let's focus on a couple of stats. First, let's look at our on-time delivery rate, which is 53% in this last year. That's down from 59% the year prior. The on-time delivery rate was impacted. We had a strategic shift in midyear into what we call engineering excellence. This is higher quality, more stable code everywhere. This focus did yield critical wins for us. We now sit on significant increases in studio stability and of course, the record-breaking platform concurrency we've all seen.

But this quality bar does mean that we've had to delay releases. An example of this was last week's glTF export, where we needed to spend more time making it more reliable, ensuring a compatibility with a lot of our back-end systems. And we just firmly believe that this is our way going forward. Prioritizing stability is one of the best ways we think we can respect this community.

Yes. Another observation, it looks like we delayed about 31% of the items on our road map versus 25% last year.

Yes, that's right. And look, we just have to apologize for that. There's an enormous number of opportunities everywhere and new things, as you can imagine, are emerging as we grow. And that said, this coming up year, we really need to focus on doing fewer things with higher consistency, and we're always going to prefer to deliver higher quality features instead of rushing them and giving you all transparency along the way.

So that's right. We're going to keep iterating. Please continue to provide feedback on what you see us doing and what you see us not doing, and we'll continuously work to improve. All right. That's enough review. Let's get on to the good stuff, starting with our vision to enable the creation of anything anywhere by anyone. We're well on our way, and our strategy forward rests on 4 pillars. Number one, we're committed to providing a stable, high-performance, universally accessible foundation for all of our creators and our users. Number two, creator productivity. We're enhancing your efficiency with superior tools for avatars, animation, seamless iteration and asset management. Number three, we're going to provide you with better tools and insights to grow. We're giving you data and AI-powered tools to make more informed decisions, iterate effectively in production and reach a global audience.

And number four, a powerful engine. We have an amazing engine, and we are evolving it to support more complex and expressive content at a massive scale, unlocking the next generation of experiences. Together, these pillars will make this platform, which is already the most powerful and flexible platform to create on even better. And we have tons of great updates on each one. I'm excited to show you what we're building, starting with inviting up Josh Anon and Taylor Russ to talk about our rocksolid platform.

Hi. I'm Josh Anon, a product manager on the Engine team.

And I'm Taylor, an engineering lead on the engine team.

One of Roblox' unique value propositions is that you write your experience once, and we make sure it runs well on millions of devices worldwide from high-end gaming PCs to low-end budget phones. This lets you spend your time creating instead of thinking about things like GPU drivers. Now there's a lot that goes into delivering this. Let's look at how we're doing. First, stability. Stability means not crashing. Our goal is to never crash. In the past year, we have significantly improved our stability. We're seeing a 66% improvement in play time per crash for iOS users. Android users have seen a 26% improvement.

One of the key ways we're doing this is by continuing to invest in how we manage memory. In the past year, we've built an entirely new memory management system internally that's rolled out almost everywhere globally. Now what's extra impressive is that this system takes experiences that used to be unplayable and crash immediately on low-end devices, and it makes them very stable and playable like you see in this video.

On Android alone, based on some of our early testing, we think the system could deliver an additional 356 million hours of annualized play time. That's a pretty big deal, and you all get it transparently for free. Stability doesn't just matter when you're playing, it matters when you're building. We've been spending a lot of time working on studio stability as well. Since last year, we've improved studio hours for failure by 90%. Hopefully, some of you noticed the difference. Go ahead. [ A couple of you].

Stability goes hand-in-hand with raw performance, and we're seeing great strides here, too. Our goal is to build systems where you fall into a pit of success and we take care of performance for you, ensuring your content runs fast everywhere. We're seeing some incredible improvements with our cloud-based LOD system that Dave talked about during the keynote. And we'll go into more detail when we discuss the future of the engine, but we wanted to highlight 2 other improvements to share with you all.

We continue to comb through our code and work on perf and push on it wherever we can. As we worked on the new LOD system, we found that mesh loading was capped per frame. We uncapped it. We saw crazy gains in performance, especially on higher-end machines. In many experiences, we saw load times cut in half or more. You all receive the improvement automatically without having to republish. Performance isn't just about the client. We pay close attention to our server frame rate as well. And we call this heartbeat. We target ensuring that our servers run at 60 hertz globally. Late last year, we started looking closely at the deep internals of how our engine interacts with the Linux kernel the server runs with. We found an opportunity with virtual memory to physical memory mapping to use something called Transparent HugePages, and that improves that mapping efficiency.

We deployed THP over the course of a month, and it had a huge impact. You can see the impact in this chart, where in 1 day, we went from a PO5 heartbeat below 45 to regularly above 50. And for our PO5 users, we're not consistently hitting 60 hertz yet. We still have work to do, but this is big progress. And of course, this all happened transparently. You did nothing, Roblox got faster and all your experiences benefited.

Next up, platform reliability. On Roblox, we take care of the server stuff for you. This year, as we've broken industry records, we've had teams working harder than ever to make sure our servers are reliable and that you can focus on growing record-setting content, not managing services. Since last year, we continued to improve the uptime of Roblox itself. Our uptime was 99.972%. Our improvements saved over 12 million engagement hours. We'll keep pushing to 4 9s and then 5 9s. So we operate data centers around the world, including our new Brazil data center. And internally, we forecast growth in supply and demand and make sure we can deliver enough compute. But recently, we've seen record concurrent users, which required more compute power than we had in our entire global grid. I call this our holy moly -- sorry. We actually had a weekend earlier this year where we ran out of compute for a few minutes.

Rather than throwing up an air saying Roblox is full and stopping people from playing, we worked around the clock so that within a few weeks, we could use third-party cloud services to run our servers too so that we wouldn't run out of capacity again. Your perspective, things just work invisibly. Roblox could handle record-setting demand, no problem. A few weeks ago, we had over 45 million concurrent players, which this investment let us handle. I don't think anybody else could do that. It's pretty cool.

That's for our infrastructure team. Finally, let's talk about friction, and let's start with everyone's favorite source of friction, bugs. We've made addressing your dev forum issues a company-wide priority. Since January, we've closed more than 4,100 bugs, which is about as twice as many as we had closed last year by this time, and there's still 3 months left in the year. A major source of friction for players isn't just bugs, though. It's cheating. People who make -- who cheat make playing together less fun. We are making it our mission to end cheating on Roblox. A few years ago, we took a step forward in Windows with Hyperion anti-cheat technology, and we're continuing to improve that technology. This year, we're focused on Android as well.

But we've also made a major change in how we handle cheating on the platform. Historically, whenever we detected a cheater, we'd make a note and run a periodic manual band wave. Now we have an automated system that runs 24/7 so that if we detect cheating, you will automatically receive progressive consequences going from a warning all the way to a permanent ban. Our goal is to reform people and get them playing fairly, which is why we have these progressive actions. Hopefully, everyone here has only felt the benefits and not experienced any of those consequences.

Since April, we've seen an 84% reduction in playtime exposure to cheaters, thanks to this system. We'll continue to invest in securing Roblox by both securing Roblox itself and by giving you tools like Server Authority. Come to the competitive games breakout talk tomorrow. We're proud of results that we've achieved this year, and we hope that you appreciate them. And we look forward to sharing more with you next year. We're going to hand it off to Oren and Eric next to talk to you about creator productivity.

Hi, I'm Eric Yip, Engineering Lead for Roblox Studio. And today, we're thrilled to share with you our latest update to boost your productivity.

And hi, I'm Oren Jacob, Engineering Lead for Avatar. We're powering up you across 4 areas in our talk today, cutting out busy work so you can innovate faster and helping you scale your team's impact. Let's start with Avatars and animation. To improve the quality of Avatars with Avatar Auto Setup, back in March, we were running at about a 47% success rate. Today, that's at 90% -- this lets you focus on creative intent and bringing your models to life. And you can leave things like rigging, skinning, caging for layer clothing and positioning for R15, attach points for accessories and other stuff to us. Let's roll the video. On the creator side?

On the creator side here is a builder in blender, you're focusing on your vision, unique model with custom pants and shirts and hair and shoes. Avatar Auto Setup then takes over. The heavy lifting is done in the background and the complex conversion process is effortless. And there you have it. As you cycle through the results, you can preview them immediately, a full Roblox-ready avatar. So come by the studio booth to check out the latest and see Avatar Auto setup for yourself.

Now let's take a look at animation. We're making big strides in Avatar animation technology today, introducing the adaptive animation format. This supports retargeting of animations across all types of humanoid rigs from R6 to R15 to complex custom bone-based characters across the board. You can author and preserve contact-rich interactions across scenes and across characters. With the adaptive animation format, animations can be created once and they'll run on all avatars that Roblox supports today and that Roblox will support tomorrow.

Yes. There we go. Today, animation blending is rather limited. It takes about 1,000 lines of Luau's scripting and a Luau scriptor to make it all work. Let's roll the video. Soon, we'll be releasing the animation graph editor, an API and studio tooling that will enable next-level character animation quality and scripting, making scripting optional to deliver on that. So studio tooling will support both visual compositing and the button of composed animations as well. This will allow all animators to create rich lifelike animations with no code.

All right. Let's jump into how studio is going to work better with your favorite tools. We know many of you prefer your own code editors. This is why last year, we released an early preview for sync, and now we're rolling out another update, folder- level syncing.

Let's see this in action. Select a data model location that you want to sync from, then a folder that you want to sync it to and just like that, you can manage your entire code base from your favorite editor, create, rename and [delease] scripts, all with studio perfectly in sync.

All right. We're excited for this one. So this is coming soon in beta, and it's going to be GA by the end of this year. All right, we also know that when you're working with tools like Blender, you often need to move your 3D models out of studio. And historically, you use OBJ for that, but it's limited, right?

Single meshes, basic texture, you're looking for more. This is why we've launched GLTF exports for the assets that you have permissions for. Now on the left, the avatar is showing multiple meshes, proper rigging and skidding intact. And on the right, 2 helmets, the OBJ looks flat, while the GLTF version shows the full PVR details with the look that you intended. This is already launched, so give it a shot, and you can always come by the studio booths to give us any feedback.

Now that you exported it, getting those models back in the studio should be easy. This is why we're introducing model [Re-important] Studio. So I know the name speaks for itself, but you know what, let's have a demo. We had a house that we imported from Blender, but we want to add some rocks to it. Now the [Re-important] Studio is just one click, no dialogues. All right?

We're going to tweak some of these rock colors. And just for fun, we're going to attach this boat. Back in Blender, we're going to update this one wood in dock a little bit. And now I'm going to use a keyboard shortcut for the [reimport]. And look, everything is intact, including that boat. The magic here is that the reimport is non-disruptive, meaning your properties and constraints stay intact, keeping you in the creative flow. This is coming soon as a beta, but you can also come by a booth to check it out today.

All right. Let's get into how to improve and organize your data at scale. More creators are using more assets this year. 50% of top creators by time spent have more than 2,300 assets and the top 5 have more than 30,000 assets each, double what we had last year. To help you manage your assets at scale, we launched our beta for our revamped asset manager, so you can easily find your assets with powerful search, sort and filtering. You can search all your assets in your users and group's accounts with partial or exact matches. You can also find your models by searching using descriptive words as well.

Next, to help you organize your assets, we'll be launching user creator folders next year. Over the past year, we rolled out several improvements to asset privacy and our sharing system. Let's go to the video. You can now share models, images, meshes and even animations with your connections and your groups.

Next, by the end of the year, we're also launching the beta, enabling you to make your newly created meshes and images private, helping you control the usage of your assets and prevent asset leaks, not that has ever happened to us awkward. We also heard your feedback that you need better tools to manage different versions of your work. This is why we're revamping the place version history View in Studio.

You'll soon be able to mark specific milestones in place version history as you iterate over placed versions and be able to easily find and navigate across them. Let's look at the video. So you can leave a note on a version so that you can document important details for yourself or your team. You can check out who and your team contributed to changes made in each version. And if you need to find a specific version, you can easily search by contributor, text or date or filter for published or manually save versions.

These improvements are all coming by the end of the year. And speaking of versioning, we also know that a lack of versioning for assets slows down your iterative workflow and causes too much inventory clutter. So next year, we'll be extending versioning to meshes, and images and improving version control for animations as well.

Your player data is crucial. And this past year, we launched the Data Storage Manager, a powerful solution for your game tester ops and community management teams and members to view and understand exactly how your player data is being stored. So let's go straight to the demo.

Say you launched a coin collecting mini game and you want to see if your game is properly saving your players' coin count. Your game tester can open up the Data Storage Manager, validate the proper data structure, search for themselves and then play the game and monitor in real time how their player data is being updated. After your game is launched, your ops numbers can keep track of your -- how large your data store has grown across how many data stores themselves and how many keys.

And when your players come to your community manager with issues, perhaps data corruption that led to players losing all their coins, your community manager can easily search for that player, verify their history with versions and proceed to rectify the scenario right there.

In the near future we'll be adding deletion workflows as well as many other capabilities. You can come to the cloud services booth to chat with the team and learn more about that right after this event.

All right. Let's talk about collaboration with your teammate and even with AI on your side. Whether you're working solo or on a team, fast feedback moves projects forward. With comments in Studio, you can collaborate directly in your workspace. Let's see it in action. For instance, pin an object like this car and leave a comment, have a task for a friend, tag them directly, then quickly filter.

When you're done, resolve comments and capture your decisions. This is live today, but we're just getting started. Soon, we're going to be adding support for deep linking, so you can get to those comments even faster. Now we also know that as your team grows, control becomes critical. This means a need for safer permission and more oversight. So we're enhancing creator groups.

On the left, we've added experience level permission, so you can grant access to one experience and not your entire portfolio. And on the other side, we've also enhanced the activity tracker for you to audit actions in real time. And coming soon, we're adding support for even more events like revenue payout to help your team scale even faster.

We've also been hard at work with Assistant, and today is a big leap. Assistant will soon speak MCP. This is the protocol to let your tools talk to one another. Let's see a demo with Figma. On the left side is a leaderboard that we drew in Figma. And on the right side, Studio is connected to the Figma Dev Mode, MCP server. And I've just asked Assistant to import this leaderboard into my place. And there it is, fully built with a working scroll bar.

But the magic with Assistant is iteration, so let's spice it up with some emoji icons, see what happens. And there you go. So I want to say that Figma is just one example. Soon, Assistant MCP will connect with the entire ecosystem of MCP servers. So this is coming soon at the beta. I invite you all to come by a studio booth today to check this out.

And so thank you very much. Next up, I would like to hand it off to Peter Yang and Kyle Spence for tools and insights.

All right. Hey, everyone. I'm Peter, and I'm a Product Manager at Creator Services.

And I'm Kyle, an Engineering Manager on Creator Services.

And today, we're excited to show you all the amazing tools and insights we're building to help all of you grow your games faster.

So let's start with analytics. We've been shipping nonstop this year, including retention cohorts, insights, external link tracking and even avatar analytics. And today, I'm proud to share that over 93% of our top 1,000 creators are using analytics every month to make better decisions.

And for the 7% of you that aren't using it, give it a try and let Peter know if there's anything that you can improve.

Thanks, Kyle.

There's always room for improvement. Now one of the areas that we're really improving is insights to help you separate the signal from the noise with your metrics. And recently, we shipped new insight reports that highlight the most important metrics and actions for you to focus on.

We've also updated these reports of player feedback summaries to give you real-time feedback on what players love and want improved about your game.

And we believe that this combination of metrics and player feedback will give you a more complete picture of the right actions to take to grow your game faster. Now picture this Kyle, we just launched a new game, and I got to admit, I'm feeling a little bit nervous. I really want to get metrics and player feedback, but I'm worried that our retention KPIs might not be good enough for the discovery algorithm.

Well, Peter, I'm actually not nervous at all. And that's because we're launching -- we're going to launch experience betas later this year to let all of you test and improve your game before it reaches a wider audience. In beta, you control who can play your game. You might share it on social or in your community or you might run a sponsored ads to reach a targeted audience. So Peter, let's focus on improving our games retention before opening it up for search and discovery.

That sounds like a good plan. And speaking of retention, one of the most powerful ways to engage our players is to use events. And this summer, we added new APIs to let you invite players to your events from inside the experience. And overall, we're just seeing a lot of excitement and many of you using events to build excitement for your games. In fact, every month, we see 4.6 billion event RSVPs on the platform.

All right, Kyle. Now I'm really excited about this next feature. Do you want to talk about it?

Sure. I'm sure many of you have felt this pain. You just shipped a new update in your game and all of a sudden, you start seeing error rates spike and players complaining about issues. Today, your only option to fix this problem is to rush to deploy a new update, and that means kicking all of your players and restarting all of your game servers.

With configs, you can wrap every update in the config flag. You now have an on-off switch that will let you disable the new content or feature without having to redeploy or restart your servers.

But that's not all. You can also use configs to update in-game values in real time. I'm not sure if Brad from Adopt Me! is in the audience. But for example, let's say Adopt Me! wants to update their starter pack. This cracked egg isn't really all that is cracked up to be. And this royal egg is way more impressive than [posh] right? So if I want to update my starter pack, all I have to do is wrap it in a starter pack config. So configures are coming later this year, but I'm even more excited about this next feature.

We're also excited to announce experimentation or A/B testing to let you test new features in your game as well as your matchmaking configuration. Experiments let you measure with confidence whether that change you made actually improved your metrics. So let's say you want to test a new onboarding dungeon to see if it will improve retention. You can run an experiment that shows half of your users is a new dungeon while the other half get your current one.

A few weeks later, you see that users who received the new onboarding dungeon had over 37% increase in [D1] retention. Peter, what do you say? Should we ship this?

Well, Kyle, nothing makes product managers happier than seeing that be a significant lift. And the only thing that will make me happier is launching config and experimentation to all of you later this year. So be sure to join our analytics session tomorrow, where we're going to have real top studios walk through real examples of using analytics and experimentation to grow their games.

Earlier today, Jared spoke about some of the impressive growth we're seeing globally in every region. So let's talk about one of the biggest opportunities in growing your games. The international market continues to expand. Just last year, we launched chat translation. And since then, we've done over 400 billion translations. There are 11 billion translations from the content on the platform, and we've expanded to supporting 17 languages for automatic translation with our newest addition to Airvic this year.

And using our translation and localization features is a no-brainer to reach Robots' global audience. In fact, in a recent test, we saw that games that had all the translations on saw an average of almost 1% increase in playtime with over a 4% increase from Chinese, Japanese and Korean-speaking players who joined the game for the first time.

Now you might think that 1% is a small number, but compounded over weeks and months, that could lead to huge playtime lift for your game. So don't miss the opportunity to make your games accessible to a global audience. And if you have manual translations already, don't worry, we will always prioritize those over your auto translations.

In Dave's keynote, you saw a live demo showing Dave and Kiran interacting with an MPC. Let's dig into the details and show you how you can leverage the new audio, voice and natural language features in your experiences. First, let's set up an MPC who can dynamically speak audio with the new text to speech system.

First, we set up audio emitter and audio speech instances and then wire them together. Then we choose a voice and set some text to play a speech. There are a bunch of parameters we can play with. Let's mess with the voice and pitch.

[Presentation]

I don't know, Kyle, I think that robots is a little bit too high pitch man. I think you got to make some more robotic.

You know what, we can. Let's make this a more robotic sounding voice. Let's just add a sound modifier and change that to our text to speech and audio emitter. This all uses the existing audio APIs so you can create the exact voice that you want.

[Presentation]

That sounds a lot better. But how can I talk to this robot as a player?

Yes, good idea. Let's also set up using player voice, so we can interact with the MPC using the new speech-to-text system. We can add audio device input and audio text to speech instances and wire them together. Let's also add a script that listens to the text change property and add a function to make the robot jump.

Using what we made, we can command a robot to jump, super cool. But he still jumps when we tell them not to. Let's fix this. And while we're at it, let's give the robot a personality so we can more naturally talk to and command it around, leveraging the text-generation APIs. So using the text-generation API and some smart prompting, we can make our MPC feel more alive. Let's name it the Build Battle 5000 and tell about the apocalyptic world it's in. So Peter, when you ask our robot what his name is?

[Presentation]

It's pretty cool.

Yes.

Okay. Let's finally hook up our bot to smartly respond to us and handle voice commands. With some more prompting, now the MPC will try to correctly follow the commands based on anything we say. So now he jumps when I tell them to jump, and he doesn't when I tell them not to.

[Presentation]

Okay. But we talked about translations earlier. So what if I don't speak in English? I only speak German. Can the robot still understand me?

Yes. Sure. We can make this work. I'm sure your users speak many different languages. To quickly show this in action, let's just add a command to our MPC so I can translate what we can say. We can do this by using the real-time translation open cloud API. You just call out the language that your player is in and what language you want the content translated to. So now when we ask our MPC how to say, hello, how are you in German?

[Presentation]

Close enough. I can ask your questions, give it natural language commands and the MPC is able to do it all. We're excited to see everything you all build with these new features. Text-to-speech is launching soon with the increased limits that you've been asking for. And if you want to use it even more, you'll be able to pay for extended access later this year.

Finally, next year, we will launch the ability for you all to create and bring your own voices, so you can use that dynamic speech generation in the exact way that you want.

Our speech-to-text beta is also launching later this year. And of course, our text-generation API is available, and we're always working to improve it to make it more safe and reliable across all content maturity experiences.

Okay. So that's it for us today. And now I'd like to welcome Varun and George to talk about our powerful engine. Thank you.

Thank you so much, Peter. Hey, everyone. I'm Varun and I work on product and Engine.

And I'm George, I'm an Engineering Lead in Engine, and we're here to give you a quick update on everything we've been working on.

Both Dave and Nick touched on our audacious goal, getting to 10% of the gaming market. The only way we get there is by giving you complete freedom to bring your vision to life.

That's right. Whether it's blocky, photoreal or something in between, we want to make it all possible. Our job is to make sure you can achieve whatever you'd like.

If we break down what goes into making a great experience, it's about enabling you to build great content, allowing for expressive avatars so your players can viscerally engage with your content. And finally, powerful engine capabilities that bring both of these together.

As we dive into each of these pillars, keep in mind that the Roblox engine makes some promises to both players and creators. These 5 promises place unique constraints on the engine, requiring innovative solutions even for the simplest of features.

With that, let's dive into our first pillar, giving you the freedom to build great content on Roblox. We're going to start off with individual assets. Now as Dave mentioned this morning, we are completely revamping how assets are both prepared and rendered on Roblox. Let's start with how we prepare your assets after you upload them.

Meshes and textures are the majority of all the assets you upload. Every day, you collectively upload 1.2 million new meshes and 1.6 million new images on average. Once you've uploaded all those assets, we process them into multiple optimized formats at different levels of detail, perfect for each player's device.

Okay, George, let's play a quick game. Let me ask you a question. This is an example of an optimized mesh that we generate today. Can you tell me what this is?

It's a chair, Varun. We covered it in a hearsal.

That's fair. But wouldn't it be great if this actually looked like a chair though? This optimization used to happen exactly once in studio when you first uploaded the asset, which means we had no way of improving it after the fact.

We've been working on upgrading this entire pipeline so we can continuously improve and optimize these asset representations. These are some of the new results we've been getting with our new improved simplification algorithm. So now your chairs will actually look like chairs at multiple levels of detail.

Finally, George. Okay. Now let's focus on how we render your assets on players' devices. This time, let's use textures as an example. Harmony is carefully coordinating things. So your players are always getting the best quality experience that their device can support. The colors you're seeing on screen here represent different texture resolutions that are loaded into memory.

Now on the left, you can see that previously, the system is naively loading up every single texture representation at the highest quality, even if the GPU might never even need it. And on the right, you can see Harmony is now smartly streaming in exactly the texture representations that it needs.

Now more importantly, it can also stream out the representations it doesn't need. And in the middle, you'll notice that all of this is imperceptible to the player because Harmony is smartly making these decisions based on the importance of each object in your scene.

When the engine uses texture memory more carefully, that means your player gets higher quality exactly where it matters, and you have much more headroom to achieve your creative vision. More importantly, this also eliminates massive memory strikes that usually occur when a game first launches. This is one of the leading causes of crashes on lower-end devices.

This is a complex scene from [indiscernible] one of my favorite showcase experiences on Roblox. Without Harmony, this scene requires loading roughly 3x more texture memory than with Harmony.

This new asset pipeline is launching later this year for both meshes and textures. So keep an eye out for announcements in the dev forum.

But wait, that's not all. All this effort to scale down on low-end devices also means that we can scale up on devices that can handle it. So we're very excited to announce that 4K texture rendering is coming to Roblox later this year.

That's a big one, George. That's a big one.

Okay. Now if you just had incredible mesh assets and even with 4K textures, you still would not have a very compelling experience like this one over here. Great content is not just about individual assets. You also need the freedom to build more detailed and expensive worlds for those assets to really shine. Let's take a look at some of the areas we've been investing in here to get even more freedom.

First, occlusion calling. Last year, we rolled out automatic occlusion calling, so the engine wouldn't waste resources, rendering things that you can't even see. For the next iteration, we're expanding the system so it can support terrain, avatars, particles and everything else that might be in your scene.

Now speaking of terrain, today, our terrain system is still capped in 2 key ways: scalability and quality. On the scalability front, the largest terrain you can fit in an experience on Roblox today is about 4 kilometers or about 16,000 spuds on each side. Our eventual goal is to remove that cap. But first, we have to optimize how terrain is stored and represented in the engine.

With some of the improvements in the pipeline, we have already been able to push things to 16 kilometers on each side. That's an area more than twice the size of San Francisco running inside Roblox.

And also, as you can see here, on the quality front, we're also improving things. Materials and lighting on terrain are much higher quality and find details and shapes on your terrain are more accurately represented. Now even with more scalable and higher quality terrain, you still need a way to populate these worlds with a lot more detail.

To help with that, we introduced instant screening a few years ago. This allowed devices to dynamically stream in and out parts of your world as necessary. Now we've seen steady adoption of instant streaming, but there were still a few pain points that we needed to address.

Yes, we heard from you that instant streaming did not work well for your experiences all the time because the radius always followed the player. With the recently announced multiple streaming, you can now locations for scenarios like binoculars or when you needed to teleport players to different locations.

Next, while streaming unlocked the ability to showcase massive worlds, it was still miserably hard to author large worlds and test things out in studio. Here, you can see an early preview of instant streaming now enabled in studio. I can be fully productive in one part of the place without having to load up the entire map. And as I move through the place, I can see the instances stream in and out in the explorer as well as around me, allowing me to quickly jump back in and make an edit whenever or wherever I might need to be.

This is the same map that Dave showed you this morning. Without this internal build of instant streaming working inside studio, it would have been nearly impossible to build that demo.

Finally, instant streaming always traded off visual fidelity for performance for objects that were streamed out. To address this, this morning, we introduced Slim. With Slim, we will now automatically generate a lightweight but high-quality representation of each model that's streamed out, perfect for low-end devices. Let's break down how this works using the same demo that Dave showed us earlier this morning.

Okay. So on the left is the original content as authored. A low-end Android device would immediately crash if it tried to render all of these triangles and instances at once. On the right, you can see the effect Slim has. Now the device only needs to render a fraction of the triangles that were originally authored, but the user gets exactly what the creator intended at the highest quality that their device can support.

If you've tried instant streaming, you might already be aware of model's LOD's limitations when it comes to visual fidelity. Here's a quick comparison of the same demo running with both model LOD and with Slim. For the same scene, Slim gives you much better visual quality for the comparable performance.

We're excited to announce that model LOD will be replaced by Slim later this year. Our vision is to also bring Slim to Studio so we can improve your creation workflows as well. And you can sign up for early access preview of Slim today at the engine booth or on the Dev Forum online.

Okay. Now let's move on to the second pillar, Expressive Avatars. As Dave shared earlier today, we are pushing the boundaries of what avatars will look like and how personalized they can be on Roblox. Just like all other content, we want to give everyone on the platform complete freedom to express themselves. We expect to add increasing avatar complexity in the future, up to and even exceeding the quality that you're seeing here.

Now I know what you guys are thinking. Done naively, this would come with an unacceptable resource tax on you, the creator. The more complexity we add and the more avatars join your experiences at once, it's going to be harder for you to express your vision. Instead, we need to ensure that avatars scale like everything else on the platform, adding negligible and predictable cost to your experiences.

To get to this goal, we're investing in 3 areas. First, all the asset streaming efforts we talked about previously will support skinned avatar meshes and textures as well. Let's take a look at the difference texture streaming can make on an avatar. Here, we're going to simulate downloading all of this avatar's textures on a low bandwidth 128,000 connection.

On the left, we have current behavior. And on the right, we have asset streaming enabled. Ready, this is going to go fast.

Let's go, 3, 2, 1 go.

Okay. With texture streaming, avatars are visible much faster and use less bandwidth. Players can feel like themselves almost immediately. And over time, their textures get crisper if their device can support it.

I think the other one just finished. Next, avatars in the distance will also be converted into the same lightweight composite models that we showed you earlier, but these Slim models will also support animations.

Finally, we're targeting specific use cases where avatars cause resource spikes when they have layered clothing on. Here, you can see a test environment with 420 avatars trying to equip a tool at the same time. This causes a resource spike, which in turn caused the tools to take a long time to equip.

That's no good.

Yes. So let's take a look at the same scene again with some of the improvements we're working on. Ready?

Again, 3, 2, 1.

There you go. So now the avatars have their tools equipped much faster. We're launching these improvements later this year.

Okay. Moving on to personalization now. Technologies like layered clothing, dynamic heads, accessories and in-experience creation have all unlocked brand-new ways for players to express themselves in Roblox, but we are continuing to push the boundaries here as well.

We recently announced UGC emotes, so you can now create, publish and list emotes in the Avatar marketplace. We also recently launched support for physically based rendering for accessories. We had to be careful to make sure that these higher-quality accessories worked well across all devices and experiences.

And soon, we'll be expanding in-experience avatar creation to also include clothing and accessories. This means you can now allow players to customize these right within your experience. And when they're done, they can purchase and equip the item right there and then use them all across Roblox.

Okay. Now let's switch gears a bit and talk about engine capabilities. These are some of the new capabilities that unlock whole new functionality and give your experiences superpowers.

In May, we announced that acoustic simulation is available in Studio beta. When enabled, sounds reflect, then around corners and are included by the environment. Let's have a listen.

When I'm inside the room, I can hear the sound reverberate off of the stone walls. And as I slowly walk outside, the sound type fracs around the corner and comes from the left As I move behind the building, the sounds become muffled just as you'd expect. Acoustic simulation will be available for live experiences by the end of the year, and we'll also be adding support for sound instances, so it works with all of your existing content.

That's awesome.

Make sure you visit the creating immersive audio workshop during lunch tomorrow to learn more.

Now let's talk about physics. Last year, we launched aerodynamic forces and engine, so you can take advantage of a whole new class of forces across your experiences. This year, we completely revamping herodynamics work, so parts can react realistically to buoyancy, lift and drag forces just like this hydrofoil. And yes, these propellers are really simulated.

All of the work that we put into realistically simulating fluid forces for aerodynamics last year will now be available for water this year.

Okay. Let's talk about server authority now. One of the most difficult multiplayer experiences to build are competitive sports games. Due to the realities of network latency and bandwidth constraints, a soccer game like this one is incredibly difficult to get right on any platform. We're working hard to enable server authority natively in Roblox. So a whole new genre of seriously competitive games like this one will soon become possible.

Yes. Instead of each player's device deciding what happened, each client makes predictions, but only the server makes the final call. If you've tried racing game before, you might be familiar with the pit maneuver, which is notoriously difficult to achieve well in a multiplayer game. This is what usually happens.

That's no good, George.

I know.

I think we can do better than that.

So once Server Authority launches on Roblox, you'll soon be able to get this.

Competitive social games like shooters, racing games and sports are such a good fit for the Roblox ecosystem that we can't wait to get this into your hands. So starting today, you can sign up for Server Authority early access preview by visiting the Dev Forum or seeing us at the engine booth. You should also stop by the creator [track] talk tomorrow to learn more about it.

Okay. Finally, let's talk about solid modeling. Over the last year, we've made a whole host of performance and scalability improvements to the inexperience solid modeling APIs like Union, Intersect and Union, Intersect and Subtract. This has enabled brand-new emergent possibilities like this car workshop where you can literally create your own tires from scratch using traditional manufacturing techniques.

This was previously limited to primitive parts. But we've been working on adding support for meshes this year. So this means that you will soon be able to perform the same operations on fully textured meshes like the container and the [no]. And we're not stopping there. We're adding new APIs like Fracture and suite to help with some common scenarios that you always want. All of this is also available in an early access preview just like Slim and Server Authority today. So make sure you stop by the engine booth or visit the Dev forum post to sign up.

We touched on a lot today, asset streaming, slim and terrain improvements to unlock the content you build, avatar performance and personalizations to unlock more expressive avatars, and Server Authority, hydrodynamics, solid modeling and even acoustic simulation to unlock more gaming possibilities. What would you build with all this cool new tech?

George, the world is my oyster now. I want to build something like a competitive game with destruction that actually affects aerodynamics, all said in a vast and detailed open world, and everything with 4K text, obviously.

Okay. Well, I can't wait to play that. That's all the time we have today. Varun and I are so excited to meet all of you at the booths and at the other events.

But wait, we have one more little surprise for you before we pass it back to Nick. It's a video. Have a great RDC.

[Presentation]

Wow, that was super cool. So now you've seen 3 things. One, you've seen how all these pillars come together. Two, you've seen that I may have been one of the people asking for emissive maps. And now that I'll be able to get my fix on Roblox, I think I can take this thing off awkwardly. And number three, you've seen how this platform is and will continue to expand the possibilities of creation for all of you over the coming months with a strong focus on core performance, amazing productivity enhancements so that you can build sophisticated experiences, more advanced tools for data analysis and experimentation and an increasingly powerful engine.

I get excited just thinking about the potential ahead. And now as we bring this keynote to a close, I want to take a step back from all the features that we showed and reflect because beyond any single tool, what really matters is the strength and dynamism of this entire community. And as I reflected on the last year, I'm marveled at how the Roblox ecosystem has become even more dynamic and open where new talent and innovators have immense opportunity to break through.

And we see this in our data. 28 of the top 100 experiences on our platform are new this year and 96% of top earning avatar items were created in just the last 12 months. This is not a static platform. It's the entrepreneurial dream in action where the right creation can shoot to the top. And our healthy ecosystem is no surprise when you see how we all collaborate with each other at community events on our forums and in our AMAs.

The recent Community Run Inspire 2025 was a great example. It was a massive success with over 50,000 RSVPs and more than 57,000 viewers.

The event featured more than 30 speakers and its game jam produced more than 400 new experiences. On our Dev Forum, we've seen more than 35,000 help and feedback posts since the last RDC. And our AMAs have pulled in more than 141,000 views with more than 150 questions answered. All of this makes it extremely clear that this is the greatest game creation community the world has ever seen. And a quick public service announcement.

Right after this, we're going to have a 15-minute break. Then we're going to launch right into our safety Q&A. Afterwards, you can attend creator-led Lightning talks, featuring guest speakers from TikTok and YouTube with a meet and greet afterwards. And for the next 2 days, I really encourage you to stop by some of our booths. They're staffed by Roblox engineers, product managers and domain experts, give us your feedback and get hands-on technical advice.

And lastly, tomorrow's schedule includes a networking event and dozens of sessions on a range of topics from performance to competitive design and intellectual property. And that wraps up our creator keynote. I want to thank all of you for your time, and I want to leave you with one final thought. The tools and features we've shared today are more than just code. They're the building blocks of the future, a future you will create. We provide the platform and you provide the passion, the creativity and the vision. You are the architects of the next generation of play, and the future is something we will build together. Let's go build it.