HIVE Blockchain Technologies Ltd Stock

Hello, and welcome to today's webcast covering HIVE Digital Technologies financial results for fiscal Q4 and full year 2026. My name is Nathan Fast, Director of Marketing and Branding at HIVE and I'll be your moderator for today's call.

Before we get started on Slide 2, I would like to briefly note the disclosures for today's presentation. Except for statements of historical fact, this presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995, such as expects, believes and similar expressions identify these statements. Actual results could differ materially, and we disclaim any obligation to update them except as required by law. For a full discussion of risk factors, please refer to our most recent SEC filings at sec.gov.

In addition to discussing results that are calculated in accordance with GAAP, we will also reference certain non-GAAP financial measures, including adjusted EBITDA, adjusted net income and free cash flow. Management uses these metrics to evaluate operating performance and believes they provide investors with additional insight and they're presented for supplemental purposes only and should not be considered in isolation from GAAP results.

Reconciliations to the nearest GAAP measures are included in the appendix to this presentation and in the press release and Form 8-K furnished to the SEC. On the next slide, I'm pleased to introduce today's presenters. Frank Holmes, Executive Chairman; Aydin Kilic, President and CEO; and Darcy Daubaras, Chief Financial Officer. I'd now like to hand the presentation over to Mr. Frank Holmes for a macro recap of the quarter. Frank?

Good day, everyone. I'm Frank Holmes, the Co-Founder and Executive Chairman that plays a role as a strategist at a macro level for HIVE. And I'm going to give you as also my other job is a Chief Investment Officer, a macro recap of what I see in this realm of data center build-out and the as demand that's taking place for AI factories, but also the collateral that we're seeing the collateral ramifications are the demand for copper, et cetera. So let's get going and we'll speak quickly and give you a recap for what's happened this past quarter and year.

Next. But before we get into those granular details, I'd like to tell all investors that they have to be prepared for volatility and each asset class is on unique volatility and it's a nonevent for the S&P to go up or down 1% in a day and over 10 days, 3%. If it goes up more than that or down more than that, that's usually an event. That's a signal for [ country ] and selling or buying [indiscernible], as you can see, is now more volatile.

So it's one day, if we look back a year ago, it was the same as the S&P and now was expanded to 2% as a nonevent on a daily basis, and the 10 day is 5% but Bitcoin is much more when we look at 1 day is 3x greater than the S&P 500 and substantially greater when we look over a 10-day period. And as you go down and look at technology stocks and gold stocks you look at HIVE when you look at [ core weave ], you can see a pattern of companies that have more debt or companies that have this leverage to Bitcoin operations just have this greater volatility. And that means that 70% of the time is a nonevent over 10 days for HIVE to go up or down 21%.

And the same thing on a daily basis, it's a 6% ball [ core ] wave, which is a pure high-performance computing hyperscaler versus the other hyperscalers like Microsoft and AWS, which is Amazon, Azure is Microsoft. Oracle has theirs. They're embedded with another bigger technology-driven company, whereas CoreWeave's pure hyperscale and HIVE is in that transition to go into a hyperscale tailor. So you can see our volatility is greater than the bitcoin, and that provides great buying opportunities. We're especially Bitcoins down 3% then none for us to be down 6%. And usually then, it's a great bounce.

So next, please. This is the team. Aydin Kilic, our President and CEO; Craig Tavares is the President and Chief Operating Officer of BUZZ HPC, which is really the champion for our AI factories especially out of Canada. And we're now coast-to-coast across the country and we continue to expand. I'll give you more color on that. Darcy is our CFO. He's based in Vancouver, along with Aydin, [ Gabriel Ige], he's our General Counsel. He's not on Montreal. He's in Europe. [ Gabriel Amos ] is the President of Paraguay's Operations. He's also electrical engineer like Aydin is and Craig. And then we have Johanna Thornblad, who is the Country President for Sweden.

Next, please. So HIVE vision [indiscernible], it operates in 9 timezones and 5 languages, and I'm very proud that we know many of these other crypto mining companies that operate in 1 state, and they are not as efficient as our team is. And that just goes for me to share with you is that we have an exceptional team that's very efficient not only in running data centers and building data centers.

Next, please. So HIVE [indiscernible] green energy in Canada, Sweden and Paraguay. This is the most significant waterfalls in the Western Hemisphere, which has led to the largest dam in the Western Hemisphere, with like 5 miles long. It's a phenomenal piece of infrastructure. And it's a partnership with Brazil and Paraguay, generates about 14 gigawatts of electricity, half to Paraguay, half to Brazil. And for a while, Paraguay was selling a lot of electricity [indiscernible], and we're not getting paid, they're slowly starting to get some money back we come along and capture this surplus energy and they get paid every month.

And that's significant because today, we're the largest consumer of electricity in the country. And we've also hired, I think, the most engineers more than in the power utility company in the overall economic development of the country.

Next, please. HIVE's top institutional shareholders [ Invesco], [ Citadel, ] we're happy to see sites come back in, Millennium management. Two [ Sigma ] and [ Velke ] funds. Some of these are pure quant funds and other are ETFs that have directional plays in this space if it's not a technology then it's driven by other investment strategies.

Next, please. I think what's really important here is that with cancer for sure, we announced the closing of a private offering of about $115 million of 0% exchangeable senior notes due 2031 with a conversion feature, which is very unique by buying a derivative that the stock was basically double the stock price just shy of $5. Now the stock is trading at 4%, but it was at the time 2 and change.

So it was pretty significant. And what's important is that we had more than $500 million wanted and we had 24 buyers. Our name got known now to many U.S. institutions that really didn't know the unique hype story. And that's led to a big trigger in the trading volume and liquidity and price discovery has expanded with this. It's also been an important what they call a signal to institutions that we were going to accelerate our growth, like we've been saying on our AI strategy this year.

Last year was building out our Tier 1 data center of Bitcoin mining, and we increased that from 6x a HIVE to 25. This year, we're focused on this huge footprint of bringing specialty sovereign data centers in Canada and Sweden.

Next, please. So in that journey, we also increased our exposure in Canada to go in what's called the main board, the biggest stock exchange in the country where all the big banks are listed away from the venture capital based in Vancouver. And that was a great opening for us because we had the maximum amount of people allowed to show up. We had to turn back I think there was 100 people want to come from 50 places. So it was a great sign of enthusiasm for our company.

And this was a day after the holiday in Canada and the day before, our stock has a big [ par ] and with -- because of recommendations in the press release we made regarding our AI strategy in Canada. So it's all fit in very, very well the timing of it, much of it unexpected that it would come together, but we're thrilled about it for our shareholders.

Next, please. So high stock rises above his 50-day and you can see here when we announced our over $100 million convertible that it was -- it went above and as stated above, so often when you announce these things, they fall below. But no, it was a signal, I was told by smart institutions in this space that we were on a fast track for growth. And we're excited that is happening.

And then we have announcement of listing on the Toronto Stock Exchange on May summer. And then we announced the BUZZ's North Star AI factory in Toronto on May 16 was a significant home run because if by accident, we did not know that it was also showed up a very smart institutional investor in this space, in the AI but became an investor.

Next, please. And I think that it's important to recognize Leopold, [indiscernible], as the CEO was called Situational Awareness. He wrote [ seminal ] piece of paper that was a white paper of over 100 pages of what you saw in the super cycle, a fascinating background, I used to be with the open shot and so he displayed a 13-F filing and because he showed like Warren Buffet shows what he bought and sold, he bought us at 4x revenue and a couple of the other data center companies or Bitcoin mining go into transformation of AI, he sold those because they had done in a crazy run to 40 times revenue. And so we were deeply the most attractive proposition. So we're happy that he bought in those shares. We bought them, I think, on the previous month, and so that filing at the same time of announcing in Canada, our AI factory, all just happened at the same time, and it was a holiday in Canada.

Next, please. So the real race in AI is infrastructure, power land and GPU chips he wrote about this in a seminal white paper Leopold and please, if you haven't read it or recommended, [indiscernible] owns the compute owns the future, Canada needs sovereign compute to remain globally competitive. Canada is an incredible place in respect hearing was created there.

The University of Waterloo has won the software IBM annual software competition, most often being the champions there. So it has strong intellectual capital that Microsoft has always tried to hire from Waterloo. And then the University of Toronto has now become the epicenter for AI with a Nobel Prize winner. And so this future, what we're building in Canada is right between these 2 universities. And I share with you is so important because we know incentives where I'm based in San Antonio, Texas, that we have the #1 cybersecurity university.

And this started 25 years ago with 100 students now it's 10,000. And that's only led to many data centers being built to San Antonio, especially the NSA, who has the second biggest office is in San Antonio and over 3,000 employees, and they're able to tap into all these graduating with degrees in cybersecurity.

And so what you see that in Canada, it's different and what we're looking at with the cybersecurity when they come out of a couple of other universities in Toronto, which is the largest city in the country and come out of the AI and AI and cybersecurity are now becoming ubiquitous in that conversation. So we're thrilled about this opportunity of being in Canada.

Next, please. So the AI [ Gigafactory ] eventually will have 100,000 GPUs. It will be a multibillion-dollar buildout. And further to that, it will throw off billions of dollars in revenue. It's a very significant asset. It would be the equivalent taxes of 3 gigawatts of electricity. When you look in that context of the population, and GDP has always been a ratio of like 1 to 10. But it's closest to the best universities in North America, along with other schools we have in the U.S. But when we look at Canada, there are 2 premier universities.

Next, please. The other interesting part about our data centers right now, in particular, in Eastern Canada that we are in Toronto, Montreal Grand Falls in New Brunswick, we're in the most important Internet backbone, which basically goes from [ Tragodown ] to Virginia and up to Boston and north of Boston, you have the most -- the highest concentration of Internet nodes. And so that is really important for the AI for moving data, collecting data, sharing data that you need much bigger pipes. And so we're right in that triangle.

Next, please. So Canadian AI ecosystem and BUZZ AI factories. We've mentioned in previous press releases, our partnership with [indiscernible] Canada, that we are now in Winnipeg. Daily revenues improved for our HPC because Manitoba assets come on stream. So we are Montreal building out that we'll be in British Columbia soon. and will be from coast to coast in Canada and be the biggest hyperscaler really in the country by a wide margin and sovereign.

Next, please. So HIVE's BUZZ HPC partnership is Dell Computer. We have universities like Columbia University had a partnership with what we did on looking at data from Paraguay to build an AI data center in Paraguay, strategic with the platinum buyer of NVIDIA chips with [indiscernible]. So to me, it's a great view here of important relationships we've been making with universities and also with other technology companies. And I push on to the next one.

So expanding the partnerships. Craig has done a phenomenal job as seen on the far left here, the CEO of Bell Canada, the largest telecom in the country and myself, we've met with Michael Dell, a couple of times now. And so these are just other visuals to share with you that from Paraguay up to NVIDIA with Jensen, President of Paraguay, the country; Michael Dell, you name it. We are building very important relationships for growth and for HIVE, in particular, BUZZ AI.

Next, please? BUZZ HPC data centers, AI training and inference optimized for both an intensive training and real-time inference. What's happened is the -- these data centers that are doing the transformation. They are going through this new rerating. And quite often, they basically sell in the U.S. a long-term contract with a hyperscaler and they get a REIT model for the renting of their data center. What we've been doing so far is we have our own electricity. We have our own property, and we've been doing this for a while now, selling compute and getting a much higher revenue and that's the vision we have for this next couple of years.

Next, please. Toronto ranks as the third largest tale tool in North America. Next, please. University of Toronto is the intellectual center of AI. Next please. The genius of AI, artificial intelligence have developed cautiousness and could one day take over the world is Jeffrey Hinton, the PhD and Noble Prize were in 2024, the University of Toronto. So it's become an important nest of brilliant minds that we're studying under -- and been starting under [indiscernible].

Next, please. This is a visual of the team in Paraguay and behind is the largest dam in the Western Hemisphere, which is generating over 3 -- over 14 gigawatts of electricity, and we have 300 megawatts expansion in Paraguay. And our long-term vision is eventually to get to 1 gigawatt of electricity.

Next, please. One thing we've always done is education and what we've done in a school, that's a kilometer away from us half a mile had outdoor bathrooms and really quite antiquated in the centers that you would expect in America, Canada or the U.S. So we've taken up to North American standards and young kids are thrilled about it and so are we.

Next, please. So while demand is running infrastructure supply. Global AI spending is expected to reach $700 billion by 2028, and HIVE is going full speed and building that and participating in this incredible bond. Next, please. AI demand is running infrastructure supply. So we let take a look at the numbers, potential addition to global GDP due to increased productivity is $4.4 trillion. It's a 3.5x growth in AI data center demand over the next 5 years and 4-plus years.

And capital markets are just changing so rapidly, the funding of GPU chips. You have BlackRock and Blackstone creating credit funds just to lend because GPU chips have now become an asset class like cars are for lending. And you can buy car loans. And it's really quite fascinating that this year, how fast or rapidly there are other sources of capital outside of banks.

Next, please. But together collateral, we want to share with you is the money manager known for my world of gold and resources copper usage is just huge, the grid and power infrastructure build and what does that mean? Copper is the heart of the electrification boom copper is making all-time highs because demand is far outstripping supply.

Next, please. And we can see that when you go to a gigawatt data center, you're going to spend 50,000 tons on copper, not pounds of copper, but tons of copper to rewire everything. So we also see that electrical cars use more than 5x amount of copper wiring. And it's just important for investors to grasp the constraints that's happening globally for this boom in AI.

Next, please. Copper demand is projected to rise 40 to 50% by 2040, where is it going to come from? It's going to come from Africa and South America. And now you're seeing all of a sudden, copper deposits being refurbished and looked at that were shut down for low grade, all of a sudden are going to be commercially more attractive in Canada and the U.S. So there is this boom and it's a collateral boom.

We've also seen fiber optics because you need dark fiber prices have doubled 100% up. So the inflation on the supply and a lot of times getting HVAC or the electrical air conditioning you need for these high-performance computer data centers they have gone from being 20 weeks to 30 weeks to 40 weeks to 50 weeks and sometimes now 60 weeks to be able to get the equipment to build the data center, which we then put your GPU chips into.

Next, please. But we still have a mix market is [indiscernible], level crisis against the AI sector regularly saying what they're shorting and debt levels and calling it a bubble in the past 9 months. These 2 guys have PhDs and [ bublology ] and running around, and this creates a market and you get sell-offs and then you get all-time highs coming back.

The demand in sales and when we take a look at NVIDIA, we're seeing that other competitors are coming in to come up with their own high-performance chips because NVIDIA is such so much further invest than anyone else and the demand for video chips are so great. So when these guys come out and talk about how negative, et cetera, in our social media and if they can surround an impact to short-term sell-off, it's usually a good buy. Next, please. we're very excited about where we are in that traction. So I now like to turn it over to Aydin Kilic, our CEO and President.

Thank you, Frank. That was an excellent macro recap. What an exciting time for us. People tell me I need to smile more because all the amazing things we accomplished this year. So here's a photo of me smiling. So let's jump into it.

So for those of you that may be new to the HIVE, we are a vertically integrated data center builder and operator. What this means, we land bank, we buy land by substations. We build data centers from the ground up. We operate data centers. We'll acquire old data centers and retrofit them. And of course, we orchestrate compute, we're an NVIDIA cloud partner, we built the BUZZ Cloud, which has been ranked by cluster mix and semi analysis. And we've got a 440 megawatts of Bitcoin mining capacity as part of our dual-engine strategy globally.

But really, we think that selling tokens and being at the forefront of the AI economies where our megawatts will get those AI multiples. So it's a very exciting time to be a high shareholder. And this photo is actual of our executive team at the [ Itaipu ] Dam, in Paraguay last year when we're building out our 300 megawatts there, sites -- sites not too far from here. So again, we travel the world, we're boots on the ground. We helicopters, planes, you name it too. The site visits, conferences, see us around look out for the gold [indiscernible] B, of course, and let's jump into it.

Next slide. So it was a phenomenal year. I think we really knocked out of the park. If you look at the business overall, we did approximately $300 million in revenue globally. We had over $100 million of gross operating margin and over $75 million in the operating income. Net operating income is our gross operating margin less corporate G&A. So in a cash business what we produce and $73 million of adjusted EBITDA. Now the net loss is booked at $148 million approximately, but that includes a very substantial depreciation and noncash adjustments.

So if you back those out, of course, it would be in the positive, but we have an aggressive depreciation schedule 2-year straight-line depreciation for ASIC in 3 years for GPUs. So just to keep that in mind. But on an ROIC basis, a solid year, 13.3%. And we've also been deploying and selling our Bitcoin to fund operations and fund growth. So you'll see that we have a modest but healthy 150 Bitcoin in the treasury as of March 31.

Next slide. Solid quarter as well, $72 million in revenue for the quarter, $17.5 million gross operating margin, and the business still did $8 million of net operating income, which I think is very admirable because as we've been growing the business, and I'm going to talk about our growth shortly. We've brought on key team members, and we have contractors and we have tax specialists, we operate in Paraguay in Sweden, Canada. We earn revenue through Bermuda for tax efficiency and how does that all fit together?

So our corporate G&A we have the lowest G&A as a function of revenue amongst the loss in the entire industry. So we still are growing the business, but we're still profitable quarter-over-quarter on a cash basis. And I think that's very important to highlight. We don't just go higher hundreds of people and burn a bunch of cash and say, "Yes, I don't worry, we'll figure it out later."

We've intentionally scaled the business for getting to critical mass and we've earned money along the way. Again, you're going to see that net loss really as a function of depreciation and noncash items. And so we always do like to point that out. Still a healthy quarter. You see that our HPC AI revenue is trending up above 5% now, so about 6%, 7%, and that will continue to grow as 100% of our growth this year is on the HPC and AI business. Next slide.

Year-over-year, as is our fiscal year end March 31, looking at that gross operating margin, I think it's very admirable. We did over 4x growth year-over-year. the $107 million of gross operating margin approximately for the year. That's up from $25 million the previous fiscal year. And of course, you see on a quarterly basis, you see it rallied as we had some really strong performance in the Bitcoin mining business fiscal Q2 last year. But overall, it's been a tremendous year of growth for us.

Let's go to the next slide. So on a net operating income basis. I think it's even more impressive because now that we're operating at scale, again, we've been making key executive hires. We've been bringing on consultants and contractors as we tactically and very strategically scaled throughout Canada, our partnership with Bell building the 300 megawatts in Paraguay and having all the tax and accounting in place to really have a truly multinational organization.

But I point to sell because even after our growth in the size of the team and the G&A, our net operating income, which is, again, gross margin minus corporate G&A, $76 million for the year, up from $8.5 million in the fiscal year before, that's 9x growth year-over-year, which I think is tremendous. And so again, when I say we've intentionally scaled it means we're paying attention not just where we're going, but what are we doing right now.

And so I think that it's going to be a really, really exciting year for HIVE because we built a tremendous machine. We attract, in my opinion, the best of the best. We have a high perimeter culture. We study and implement the teachings of [ Jim Collins], famous author [indiscernible] great. And I think there's going to be some really phenomenal success in the year ahead.

Let's jump into the next slide. So again, our dual-engine strategy, the cash flow from the Bitcoin mining business allows us to scale and grow the more long-term and stable HPC colocation revenue and GPU cloud revenue for a BUZZ. Now this is a snapshot of where we finished the fiscal year, March 31, 2026.

Now at the time, the Street didn't know our revenue was going to be $300 million. This is a snapshot of our market cap at the time, aligned with where our actual revenue was. So you could see we're doing a little over 800,000 daily revenue at the time. But our market cap really pulled back to the sub-$500 million because there was a big pullback industry-wide in February. We saw Bitcoin get into the low 60,000s. We may -- and I want to say, I want to point out -- we made it through with a profitable gross operating margin and a profitability net operating income.

For this quarter, even with all that calamity in February. And again, quarter-over-quarter, for 6 years running now, we have mined with a positive mining margin quarter-over-quarter. And we downfall, we optimize, we curtail. Pound for pound, I believe we are the best bit clan miner in the vision. We know that The Street is very much focused on HPC. But when you have that cash flow engine of Bitcoin mining, that's funding the growth, we want to make sure that, that engine is a very well-oiled machine in pound for pound, best-in-class.

So 25 exahash installed about 23 exahash average operational for the quarter, 876 Bitcoin mine. BUZZ about $5 million revenue for the quarter, $35 million, which would be $20 million ARR, but contracted $35 million because we had this exciting Blackwell deal. We'll talk more about that we brought online. In our target -- our target was $200 million ARR for GPU cloud business and about $300 million ARR included the HPC colo capacity that we had as well.

Let's go to the next slide. and here we are today. So revenues jumped up. We're doing about $350 million ARR, a little over $900,000 a day. And this is as of June 1, Bitcoin's at $71,000. This was closer to $1 million a day a few weeks ago, but that's okay. The Street is starting to pay attention, $1.2 billion market cap.

Well, we're at the catalyst there. We did that phenomenal $115 million convertible bond at 0% interest and that was a massive catalyst to fund the growth of our GPU Cloud to double that GPU cloud from 5,500 to 11,000 GPs and realize a $200 million ARR. And then, of course, huge news was our Toronto area Gigafactory, which increased that ARR target for a collective HPC business to $660 million. So it's a $200 million ARR for the GPU Cloud, plus $440 million, if you look at HPC colo.

So now it's a really exciting time. The stock is actually as of today, June 1, we hit $5. So getting into that nice institutional range and, of course, being over $1 billion market cap U.S. So it's really great to see how -- it's important for us, we realize having these targets, but also showing our growth capital which we always target lowest cost of capital, of course, to realize these numbers, and it's just a good feeling for our shareholders when the market rewards us for this being astute stewards of capital.

Let's go to the next slide. So again, just again, it's our year-over-year. So just a quick by the numbers. Our cash rate grew over 200% in operational cash rate for the fiscal year. And if you look at what was installed as of the end of the fiscal year grew almost 20%. Again, with Bitcoin at 71,000 today, mining over 11 big in a day, it's about $100,000 baseline revenue and be very intentional about how we scale.

Let's go to the next slide. I do want to point out, again, navigating the volatility implicitly in the Bitcoin mining sector. We've talked about hitting our 25 exahash we did that. But what we then did is we optimize firmware for all the different types of machines that we had. And we got that hash rate, it's actually 24.6%, although on an installed stock basis, over 25 exahash, but the trade-off is the efficiencies improved 16 jewels of terahash. What that means it lowers your breakeven cost of mining.

So your total output cash rate is slightly lower, but the trade-off is your breakeven cost improves. And so in bear markets, this is what you strive to do -- it's sort of -- it's a plan or mass solution that we constantly optimize. But again, it's really about having this level of expertise in the background, that cash flow engine that's helping us spur up and expand the HPC business. And even though we have 440 megawatts, we're only consumed about 395 megawatts because we brought online, more efficient machines. We've replaced some bus manager with F21 XPs, which were bought with credits we had from our [ Bitmain ] pledge last year. So very strategic intentional and curated way to make sure the business continues to cash flow through any volatility.

Let's go to the next slide. Okay. And the final slide on the Bitcoin part is really just to give the readers out there. What does this look like? When you talk about volatility, will real simple. Here's the rubric, 70,000 Bitcoin, doing about 300,000 day profit. 80,000 Bitcoin, a little over 400,000 day profit. And 90,000 Bitcoin over $500,000 a day profit. And this is an illustrative example, if you assume an electrical cost of $0.05 because based on electrical costs, well, that's the cash flow from the machines and those goal-toward pay other operating costs and of course it's corporate G&A.

But this just gives you a flavor of where the one engine, how much cash flow is producing and again, we're actually 16 jewels at Terahash with everything optimized 24.6 exahash. And now let's launch into the next section.

Perfect. So zooming out globally, 860-megawatt footprint as of today, 440 megawatts of active capacity, again, actually consuming 395 megawatts of power, but we've got 440 megawatts of data centers globally. And with the GTA Gigafactory recently announced more on that later. And of course, the Phase III [indiscernible], that brings our total roundout to 860 megawatts.

So I think it's going to be a really exciting year ahead as we provide the Street updates on how we either convert some of our existing Bitcoin mining capacity to HPC or in cases like Yguazú and our Gigafactory and the GTA develop those and bring those to market. So it's going to be a really exciting year ahead with lots of updates.

Let's go to the next slide. So focusing on the cloud, BUZZ cloud, this is our GPU business. As you know, we're doing $35 million of annualized revenue. Today, that's realized. So that's spread over the 5,500 GPUs. And so the green bubble here on the left, that's what's active you see a bit of overlap now as we grow into that Bell AI fabric partnership. So again, our partnership with Bell AI fabric is really is a colocation. So we are standing up our GPU clusters in Bell AI fabric data centers across Canada. -- currently contracted. We had Manager in British Columbia, we've press released this. What this does is it gives us a quick time to market, low CapEx path to scale our GPU cloud revenue. And so of course, you have the stamp of validation as large as telecom clear Bell Canada, choosing Buzz exclusively to build their -- to be their data center -- sorry, their GPU cluster, orchestrator and operator.

And so what does that mean? So if we have our own clients and we're standing up GPUs and AI Bell fabric data center, we're just paying the colo fee, 20% below market, very attractive. And if Bell brings us a customer, then they get a small rev share from, I think, 5% roughly. And so it's a massive demand funnel from Canadian enterprise clients that are looking for sovereign AI compute. And so it's a phenomenal partnership.

And you can see that ramp, but more specifically and what the catalyst was for our $100 million convert was upsized to $115 million in April was to fund the 2 large GPU clusters we have incoming. So let's talk about that briefly. So 2,304 GB200s, MOU signed 2,080 GB 300s/MLU signed. So million ARR to date, signing 1 of these deals gets us to $100 million ARR and the second large GPU deal gets us to $170 million in ARR.

And I think that's really exciting for 2 reasons. One, once we track $100 million ARR, I think our stock rerates again, again, we hit $5 today, and we're still -- and I think that was on the strength of the Gigafactory announcement. But I see each one of these GPU clusters adding a few hundred million dollars of enterprise value to the company once the definitives are announced. And again, now that we have the funding in place, I'll give you the numbers.

So the 2,000 GPU cluster, rough number is about $175 million. And so to get a sweet spot, if you want to get single-digit interest rate, what we're finding from blue-chip lenders, they like to see 80% LTV. So that means we come in with 20% down payment. So that would be a $35 million down payment on a $175 million cluster. Well, we would just raise $115 million. So we now have the funds to put a $35 million down payment for each one of these large cluster deals.

And what does that do? Well, again, we got the MOE. We got the data center space with Bell. We've now got finalizing the financing. So when we announced the definitives, we'll be able to advise you on total contract value at a 3-year, 4-year contract, all the particulars, I know every base curious, really, this slide is to forecast to you hey, this is what the incremental ARR will be for each cluster being online and how we get to $200 million ARR. And of course, the final tranche in Q4 is just to fill up the remaining capacity in [ Manitoba ].

Again, we currently have 500 B200s there. And so there's a pipeline for about another 1,500 B200s or 300s at that site. So collectively, that gets us to over 200 million ARR. On the lower half of this chart, are other sites we've talked about. Again, our Toronto airport site, the smaller [indiscernible]. And of course, New Brunswick, which is our flagship Bitcoin mining side in Canada, all of those converted to HPC colocation is Tier 3 data centers. And you can see what the ARR, again, on HPC colo and we're forecasting under $30 a kilowatt for New Brunswick and higher, of course, for the GTA sites.

Now the Gigafactory adds a massive $360 million ARR just on a colo basis. We expect that site to be completed and energized late 27 and active and cash flow in early 2018. So you could see now our constituent target total ARR for HPC is $660 million, $200 million from the GPU Cloud and $460 million on HPC coal, a very exciting time for a high, very exciting time for buzz and a very exciting time for our existing shareholders, and we welcome new nesters as well. So let's go to the next slide. This is a quick overview of that $115 million changeable mill.

I wanted to slot it in here because this tied into the funding of those GPU clusters. So again, was $115 million convertible note due to 2031. 0% coupon and why I think this is so great, if you did an equity financing our stock was at $2.18 when the deal was priced. So typically, bankers, The Street is going to want to see a 10% discount.

So if we did an equity financing, it would have been dilutive, it would have been probably at $2 and you're paying your 6% or 7% broker fee. This was tremendous because at 0% coupon, really what you're telling The Street is, hey, I'm doing a 0 interest bond. And if it gets exchanged, it will be actually exchanged at a premium.

So it's like doing equity finance at a premium to your stock price, not at a discount. And the great thing, and so the base conversion prices have been $2.57, we bought a capped call at 125%. What actually put our conversion premium of $4.92. So that means with a cap call, there's no dilution up to a $1.2 billion market cap. And by the way, we get those proceeds. So beyond the 257 conversion price up to 4.92%, we actually get that as a payout from the [ Capco ] stakeholders. The value of that payout look at the right hand of this chart is actually $105 million.

Now there was a cost to that cap call, which was $19.8 million. About $105 million payout for a $20 million that's over a 5x payout ratio on value. That's a good insurance policy to me. And it just so often to be today on June 1, we actually rallied past our cap call price today. So very exciting times.

Of course, that $105 million payout is based on maturity. If you do an early conversion, you might have to negotiate that. But I think that this note was phenomenal, and Cantor did a tremendous job and so we're really looking for the notes are trading well as well in the secondary market. So it was a great inaugural debut to the convertible market for HIVE. And we're deploying these proceeds to get those GPU clusters funded and definitive agreements announced in the near future. So stay tuned for big updates on that.

Next slide, please. Now I want to put a little bit of context, I want to put context on our recent [ Blackbaud ] deal. So this is the 504 GPUs. We've talked about this deal. It was in our last presentation. It's live today. It's great. It's cash flowing. It's the first Blackwell cloud in Canada, [ 5004B200s ] in that Bell Canada, Winnipeg site. But what I want to point out just for The Street and for all the listeners to date, this was a 2-year contract that was valued at $30 million for a cluster of GPs that cost $30 million.

So I'll say that again. We effectively sold the entire face valuing those GPUs upfront in a 2-year contract. Now we do have some OpEx, of course. It is being co-located in the [ Belle ] fabric site. So actually, your ROI, it's a 2-year contract for $30 million after operating cost, your ROI is more like 2.5 years, but that's still tremendous. We've signed the entire face value of these GPUs upfront. Now what does that tell you? Well, it tells you the amount of demand for these EI natives and enterprises that want -- that computer sold value valuable to them, they will pay the entire value of those GPUs upfront in a fixed contract. When does that get them to get some exclusivity, you get some sovereign compute. It also gets the white glove service from BUZZ, where we set up the GPUs in the data center, which, of course, is nontrivial and we sign an SLA, and we make sure they get that level of service and quality that they expect.

So of course, that's what we're good at. That's what our expertise at. We're masters of orchestrating compute. And again, we've got 5,500 GPs globally. We've been doing GPU Cloud for several years now, and now we're starting to hit critical mass. But I just want to point out the virtue of the GPU cloud business, we are effectively seeing deals, long-term deals where you're able to go purchase a cluster GPUs, sign a long-term offtake contract where the entire value of those GPUs and then some getting upfront in a contracted revenue.

So really excited, really bullish on the growth of our GPU cloud business. as we deploy those proceeds from or convert to bring online those 2 large clusters. And by the way, it's a CapEx-light strategy for us to scale. We, of course, have our own sites that we're going to be building for Tier 3. But in the interim, having this CapEx light ramp with Bell Canada, it's near quick time to market, low CapEx. I believe it was a great way for us to really lead the Canadian soften AI ecosystem. Very happy with this.

Let's go to the next slide. So again, circling back now. I pointed out on the world map, we have these 2 large 2,000-plus GPU cluster deals in the wings. So what I can tell you is directionally, we're very much aware that we're actually targeting 3-plus year contracts now. And what is that too? Well, if the ROI on the GPU cluster after cost might add is 2.5 years, you go ahead and sign a 3-year contract upfront.

Now you have locked in the face value of the GPUs across OpEx and you fully paid your GPUs off plus contracted profit in the 3-year term. If you did a 5-year contract, for example, now, and it's a 2.5-year ROI after cost. Now you've 2x completely paid off your GPUs upfront with a contract and you have the residual value of those GPUs at the end of the term.

So what's really interesting is we had one highly renowned institutional shareholder that is deploying a lot of capital in the AI sector, they were actually the opinion they prefer shorter-term GPU contracts of 1 in 2 years because they think that there's a lot of upside. They think you're leaving money on the table. When you sign a 3- or 4- or 5-year contract, obviously, that's done at a bigger discount and they think they're very bullish on the residual value GPUs.

Other highly respected institutional investors, more from the trade world, amongst our largest investors actually have a different opinion. And they like the stability and they say, signed a 5-year contract if you can or a 4-year and that way or 3-year and then you've at least locked in the entire value of the GPs plus some profit and whatever the residual value is just a cherry on top. But directionally, this slide is really just to share with The Street, hey, we're going long term. We're going big. And we're on our way to hitting that $200 million ARR target this year on stay-tuned for updates.

By the way, that's an actual photo of one of our deployments of [ H200 ] deployment in Quebec, Canada. So it's a very beautiful stuff, very sophisticated stuff. Let's talk to the next slide.

So really, this is just a slide to focus. If you just look at Canada, if you look at BUZZ in Canada, BUZZ HPC in Canada, this platform alone between the Gigafactory site, the smaller Toronto site and our New Brunswick saying, they have about 400 megawatts of utility load and almost 100 megawatts of -- sorry, 400 megawatts utility load in 100 acres of land that we've assembled to bring this capacity. And through 2028, this will transform into $450 million of EPC colo revenue alone just in Canada.

So I really want to prop up the strength of our sovereign offering in Canada. And I think that our Buzz HPC team, Craig and the team have done a phenomenal job and we're -- it's the first Blackwell cloud, this tremendous partnership with Bell Canada now with our GTA Giga factory, which, of course, that's a high in BUZZ owned opportunity. We'll be providing The Street a lot of updates on that.

Let's talk to the next slide. Really, this slide reminds people, we once were operating a fleet of 130,000 GPs in Sweden during the theory of mining days. So we know it's 1 or 2 about orchestrate and compute. And so for us to say we're going from 5,500 GPs to day to 11,000 GPUs to target end of year. It's a very exciting goal and I would just say stay tuned.

And just a reminder, we can do bare metal offerings with our GPUs or we can offer through our BUZZ cloud. Again, -- we've built that we have we have [ Kubernetes], we are able to sell and manage AI services, which matters a lot for enterprise clients that we may get through partnership with Bell or any other enterprises that want the full cloud offering or the bare metal. And again, we've been growing this cloud business since 2023.

Let's talk to the next slide. So the crown jewel, the slide that everybody is talking about. So really, the Gigafactory, it means 100,000 GPUs. And so the CapEx to build this will be about CAD 3.5 billion, but that will throw off 316 million ARR. And so I think that's a really exciting target to have. We expect the site to be energized by end of 2027 and live with compute in early 2028. So this is a 25-acre site in the Greater Toronto area. We spent $58 million on land.

So I've alluded and I said we've been land banking by substations. We guess what, that's exactly what we've been doing in Paraguay and throw Canada as well, even in New Brunswick. So you need land and power. Those are the 2 constituent things you need to realize and build a Tier 3 data center and bring that to market. So we've got over 90% renewable energy. And we're working on a closed-loop 0 water use design, sub-1.3 POE target. And this is going to be a great job creator for the region. So we're really excited, standby for updates. There's going to be a lot of news as we provide The Street more color on developments for this massive game changer of an opportunity.

Next slide, please. And really, this has been very intentional. This is a rendering of the conceptual slide. This is a positive impact for the community to -- and this doesn't happen overnight like -- this deal has been well over a year, and we've been engaged with the region in the municipality and even in the community. And what does that mean? Well, you don't just go in and drop in a data center. This is massive upgrades to civil infrastructure. We're talking about widening roadways, upgrading regional water line.

There's a lot of NIM, no, they're going to sap up and use all the water. Now it doesn't work like that, who are actually upgrading the water lines regionally, before, we were going to build a data center here, there was a development application in place for typical use. And they -- the region said, look, you've got to upgrade. You got to do all these civil upgrades, water line, storms, sewer, et cetera. And often site development oftentimes application get log jam when the region has -- or municipality has these large civil -- it's been almost a decade in commercial property development.

And [indiscernible] notice when you drive bio when a new high-rise goes up like all the roads leading up to it, all of a sudden are brand new and much nicer, Well, that's specific contribution that you have to make. So actually upgrading and improving the community -- it's going to create hundreds of skilled jobs. And again, it's based on closed loop liquid cooling.

So deeply entrenched into the wants and needs of the community, very thoughtful design, satisfying both the region and the municipality. This stuff doesn't just happen overnight. We've been at it for over well over a year. And so it's very exciting that we have this announcement recently. Stay tuned for more updates.

Next slide. So I'm going to call this a new era for HIVE. I see a $5 billion U.S. market cap on the horizon and beyond. And what I mean by that is we've got a 500-plus megawatt global HPC power pipeline. So you may recall that earlier slide, I said if you only focus on Canada or sovereign about 400-megawatt pipeline of capacity across 3 sites. And if you look at Sweden, we've got approximately 40 megawatts of capacity, our big bone and little Boden site and then, of course, the 100 megawatts in EU. So that's very exciting as well. And how do you justify that $5 billion market cap, I'll show you.

Let's go into the next slide. So if you really look at our peers, Iron obviously has done a tremendous job scaling their GPU cloud business, and they're trading at about a 6x multiple there. Now our peers that are focused [indiscernible], PLD both [ Cipher and Hut]. The interesting thing is companies like [ Wolf and Cipher ] -- sorry, [ Wolf and APL ] actually have revenue today on their HPC business. And [ Cipher and Hut ] is actually all forward contracted. They don't actually have revenue today. And so that's what we have these 2 buckets.

So Tier 3 means active current revenue in HPC and scaling and then Tier 2 is really just contracted HPC revenue. And you can kind of see the average multiples. So you're actually seeing a higher multiple in the Tier 2 bucket with only contracted revenue, about 15x enterprise value to target ARR revenue for Cipher and had compared to 10x multiple, if you look at the average PLD in Wolf. Nevertheless, we're just using a nominal 8x multiple on our 2-year forward colo revenue, which, again, really all these sites that we talked about, the Gigafactory New Brunswick, I refer to the previous slides, but they'll sort of come online through the course of '27 and then Gigafactory early '28.

And so if you line up the targeted ARR we have put an 8x multiple. Just our HPC colo business, look at the top rate is about $2.9 billion enterprise value, plus another $800 million for New Brunswick, Toronto [indiscernible]. And so that brings you about $3.6 billion. The cloud business, we actually put a 5.9% multiple on that, like iron, and that's about $1.2 billion. So if you add that up, it's about a $4.9 billion enterprise value. And then if you put a normal $500 million valuation on the Bitcoin mining business sort of using the blended valuation where [indiscernible] puts you at a $5.3 billion implied enterprise value using the sum of the part.

Downside case. Now that puts us at about 4x rerating where we are today. And again, this is, of course, predicated upon the GT gate factory having an HPC lease side and, of course, is contracting those GPU clouds. So again, state. We've got a lot of exciting updates coming as we develop these sites, and we'll be announcing contracted revenues in due course, really just shows where we would be trading amongst our peers at similar multiples.

But let's go to the next slide now. If you actually take a look at where our peers are. Again, we use a base case of 8x multiple in the colo. Some of our peers are trading well in advance of that. So at a 10x colo multiplier, using the same 5.9x on the GPU Cloud, puts us a $6.8 billion base case. And if you look at the blended average of all our peers on the colo multiple Wolf cypher, APL, hot, et cetera, the blended average for the sector items actually 1.4x 2-year forward revenue, it would put us closer to a $7.6 billion upside case.

So again, I really say it's a $5 billion and beyond outlook for HIVE right now, which is a tremendous tremendously exciting time. And standby as we continue to execute, provide treat updates on contracted revenues, install more GPUs advance our site developments, order long lead items, broadcast ready-for-service dates. All that good stuff that you'll expect seasoned Tier 3 data center builder to provide as well as updates as we expand our NVIDIA Cloud. Thank you very much. Over to you, Darcy.

Good morning, everyone. Fiscal Q4 was another productive quarter for HIVE as we continued executing on our strategy of scaling digital infrastructure while growing our HPC and AI capabilities. Overall, our results reflect continued growth in our operating platform and demonstrate the benefit of maintaining diversified revenue streams across both hash rate services and high-performance computing.

Looking here at our capital structure as of March 31, 2026, I've had approximately 259.4 million basic shares outstanding. We also had approximately 3 million warrants 2.6 million options and 15.1 million RSUs outstanding. Throughout fiscal 2026, we were able to access capital markets to support growth initiatives and strategic expansion projects. Our focus remains on allocating capital towards opportunities that we believe can generate attractive long-term returns while maintaining financial flexibility.

Going to the next slide, looking at our quarterly results. Revenue for the fourth quarter totaled $71.8 million, and we generated approximately $9 million of adjusted EBITDA. We produced 876 Bitcoin during the quarter. reflecting the continued contribution of our global hash rate services operations. While hash rate services maintain our largest revenue source today, we continue to see encouraging progress from our high-performance computing and AI business, which generated $4.6 million of revenue during the quarter. We view this business as an important long-term growth opportunity and demand for compute infrastructure continues to expand and be strong.

Taking a look at our balance sheet on the next page, it continues to be the quality of our balance sheet that is kept us strong since our inception 8 years ago. At March 31, 2026, we held $23 million of cash on hand, $10.8 million of digital currencies and $9.7 million of investments. Total current assets were approximately $59.8 million. These resources continue to provide liquidity to support operations and growth initiatives.

During the year, we maintained our disciplined approach to funding expansion while preserving capital flexibility. Although we continue investing in infrastructure and strategic growth opportunities we remain focused on maintaining a healthy balance sheet and prudent capital management.

The next slide highlights the progress we made in expanding gross operating margin dollars. Gross operating margin increased from approximately $8.8 million in the fourth quarter of fiscal 2025 to approximately $17.5 million in the fourth quarter of fiscal 2026. The increase reflects the substantial growth in revenue generated by our operating platform over the last 12 months. While market conditions continue to fluctuate, we remain focused on operational efficiency, energy optimization and disciplined cost management.

Looking at year-over-year performance, Revenue increased from approximately $31.2 million in the fourth quarter of fiscal 2025 to $71.8 million in the fourth quarter of fiscal 2020. Gross operating margin increased from $8.8 million to $17.5 million over the same period. Our gross operating margin as a percentage of revenue moved from 28% to 24% and we are encouraged by the significant growth in both revenue and gross profit dollars.

As we continue expanding our infrastructure platform, our focus remains on generating sustainable operating cash flow and long-term returns on invested capital. Comparing the fourth quarter to the immediately preceding quarter, revenue was $71.8 million compared with $93.1 million in Q3. Gross operating margin was $17.5 million compared with $32.1 million in the prior quarter. The sequential comparison reflects normal fluctuations in Bitcoin mining economics, market conditions and operational factors affecting production and revenue during the quarter.

Importantly, the business continued to generate positive gross operating margin, demonstrating the resilience of our operating platform even in an ever-changing market environment. This next slide highlights the distinction between operating performance and reported earnings.

Adjusted EBITDA improved year-over-year, moving from a loss of $30.7 million in the fourth quarter of fiscal 2025 to adjusted EBITDA loss of $9 million in the fourth quarter of fiscal 2026. Reported net loss for the quarter was $76.3 million compared with a net loss of $72.9 million in the prior year period. Primary difference between these measures relates largely to noncash items as we have discussed on prior webcast, even though those charges do not impact current period liquidity.

For that reason, management continues to monitor both GAAP results and operating performance metrics such as adjusted EBITDA. This slide highlights the distinction between operating performance and reported earnings. Adjusted EBITDA improved year-over-year, moving from a loss of $30.7 million in the fourth quarter of fiscal 2025. And to adjusted EBITDA loss of $9 million in the fourth quarter of fiscal 2026.

Reported net loss for the quarter was $76.3 million compared with a net loss of $52.9 million in the prior year period. The primary difference between these measured relates to largely noncash items, including depreciation associated with our ever-growing infrastructure asset base. stock-based compensation and various accounting adjustments required under U.S. GAAP. As our infrastructure footprint expands, depreciation expense naturally increases even though those charges do not impact current period liquidity.

For that reason, management continues to monitor both GAAP results and operating performance metrics such as adjusted EBITDA. Looking sequentially on the next slide. Adjusted EBITDA was negative $9 million in the fourth quarter compared with positive 5.7 million in the third quarter. Reported net loss improved modestly from $91.3 million in the third quarter to $76.3 million in the fourth quarter.

Quarterly results can be affected by a variety of factors, including Bitcoin prices, network difficulty, production levels and accounting adjustments recognized during the period. Our focus remains on executing our long-term strategy, expanding our infrastructure platform and positioning HIVE to capitalize on opportunities across both Bitcoin mining and high-performance computing.

To conclude, fiscal 2026 was a year of substantial growth for HIVE. We increased revenue, expanded our infrastructure footprint continue building out our high-performance computing business and maintained a solid liquidity position. We believe the investments we have made over the past year strengthen our foundation for future growth and position the company to benefit from increasing demand for both digital asset and high-performance computing infrastructure. With that, I'll turn the call to Nathan running our Q&A portion for our covering analysts. Nathan?

Thank you, Darcy. That concludes the presentation for today. We'll now begin the question-and-answer portion of our call. [Operator Instructions]. Our first question comes from the line of Joe Vafi from Canaccord.

Congrats on all the progress, really exciting times here for HIVE in the industry. Maybe kind of start on the Gigafactory a little bit, great vision and plans here. You drill down a little bit more on procuring power for the Gigafactory, Obviously, power is a constraint here. And I think a lot of investors here are pretty familiar with kind of the power market in the U.S. But if you could lay out the power market in Toronto there for scaling that and in general, for Canada, that would be a good place to start. And then I'll have a follow-up.

Yes. Thanks, Joe. So the power in the [ GTA ] is governed by there is a provincial regulator. And there's actually 2 ones in charge of generation and one is in charge of transmission. And so effectively, you've got to be contracted with both and that's really the pillar of having this allocation of power. And really, that's what I could tell you for now. It's about over 90% renewable energy. And let me now is anything more specific that you'd like to know.

Just maybe a little bit on -- is [indiscernible] really a constraint there? Or I mean -- or do you see the Gigafactory kind of scaling slowly and it's really more of a CapEx spend on GPUs and infrastructure and power is less of a less of an issue? Just trying to understand that a little bit better.

No. I mean there's a full allocation for the 320 megawatts. And maybe what you're asking about is there's grid studies and load studies, et cetera, that go into an allocation of that scale. And that's what's so exciting about this site is that we have that allocation. So I think it's more what I tried to address in my section. And we've been at this for over a year now.

It's an exercise in, of course, securing the distribution and low generation contracts. And we can provide more color on the sort of detail as we provide updates on the site development, et cetera and but it's also regional and civic planning. And I had a slide dedicated to that. You don't just go drop in some modular containers like this is a real -- this is going to be a real bellwether for the community.

And so as far as your point about CapEx, I mean it would be financed really like any other large data center project. We've seen a lot of our peers issue corporate bonds. We've seen, of course, obviously, once you have lease with a hyperscaler and offtaker that obviously is a major catalyst towards funding as well. So really as we've seen a lot of our peers funds larger scale future builds. I would -- that would hope address the CapEx portion of your question. But again, please let me know if there's something more specific you want some clarity on.

That's great. That's a good backdrop. And then maybe some more color on Paraguay or I mean it feels like it's a great opportunity I know Frank mentioned plans to perhaps procure up to a gigawatt down there over time. I know you've started with some AI services for Columbia University from down there. Any other updates for now on the Paraguay opportunity and what to expect there, say, over the year?

Yes, yes. Actually, I'm glad you brought that up. It was a lot of stuff in the updates. And really, we had all this exciting news out of Canada. But it's funny you mentioned that. I promised everybody that was going to [indiscernible] the question. So we actually had an update.

Our researchers out of New York at Columbia University that we're running compute nodes out of Paraguay, their research initiative was successfully completed, and they have submitted their work, they're inaugural research using highs in tension with the researchers based in New York to the ICML. So it's an international congress for machine learning, very prestigious world renowned and Congress [indiscernible]. And so that's really exciting. And so we are going to release some of that.

I mean that's -- it's a great research project, and so I'm happy to talk about it. briefly on this call. But more importantly, is now we have -- we actually have the data on tokens per second bandwidth and latency between New York and SMS. So that was a R&D initiatives that went very well. And so stay tuned for more updates as we kind of update.

And I think one thing I've alluded to in a lot of our fireside chat, which we're always grateful to be on that you host and being at the different conferences, et cetera, is all pretty forward as I've been saying, we've been land banking by substation as well that applies in Canada as evidenced by some of our exciting recent news, but also Paraguay.

And by the way, we've been expanding our footprint in New Brunswick as well because you need more land for expanding into a Tier 3 than you do for Bitcoin mine, of course. And by the way, like when you think about it when you think of construction and you're building and you're getting into mobilization, where you're going to have storage and where you're going to have a laydown of or even just encampment for construction workers, et cetera.

So you obviously -- you need more land than just the square footage of the finished building. And you need to think of logistics as over the course of construction, et cetera. So we've been expanding our land footprint in Paraguay by the Yguazú substation, which we think has incredible long-tail value. I could say other large data center players in the industry have indicated interest. I think there was a photo floating around with [ Cruso and Pena]. So just to give you an idea of who's been poking around that area that neck of the woods. And again, we've been land banking by the substation was. So stay tuned for more updates as that initiative unfolds over the next quarter as well.

Thank you, Joe. Next, we'll go to the line of Mike Grondahl from Northland.

Hey guys, can you hear me?

Loud and clear, Mike.

I wanted to get a sense of CapEx maybe the next 12 months. Any rough estimate how you can frame that up on the BUZZ side and the Bitcoin side?

Yes, absolutely. Sorry, I thought I addressed that in my section, Mike. And really, that was captured. If you guys want to watch it on the playback if you watch the YouTube link, sort of that section that shows the growth of the BUZZ cloud in terms of hitting that $200 million ARR ramp. And then right after I talk about our $150 million 0 coupon bond. And so that sort of addressed the CapEx. So just sort of rough numbers just to give you something broad.

And you could substantiate this, I'm sure you've got similar data out there. But if you look at a cluster like what the guidance we've provided is our growth in ARR would the catalyst being the cluster size of the GPU deals for MOUs that have been signed. So we have a 288 cluster, which, by the way, when you look at [ NDL72-Grace ] Blackwell. That equates to 29 racks, so 29 x 72 GPUs is 2088 and then the other cluster are slightly larger. It's 32 racks, times 72 GPs is 2,304.

When is GB 200, where [indiscernible] 300. I'll give you a rough number. each cluster. And again, you could just substantiate this by -- or validate this by talking to peers in the industry, about $170 million for everything like when you get a [ bond ] from an NII OEM and networking your storage, everything that you need really to fully deliver is about $170 million per cluster. And so that's the CapEx per cluster in each one of those clusters, one had 70 -- sorry, $70 million ARR, the other add $65 million ARR.

And I discussed how you financed that if you want to get a single-digit interest. Typically, you're looking at putting 80% LTV, so 20% down payment. So what's 20% of $170 million will be about $34 million. And by the way, there's customer deposits involved too, which may possibly offset may not. But anyway, if you want to use a rough number, that would be a good indicative down payment from us towards each large cluster.

And then the filling out the Winnipeg site, we showed another 1,500 B200 [ 12 ], our original cluster, I also had a slide that detail the cost of the original B200 cluster, which is live now. And that cluster of 504 GPUs was about $30 million. And again, that slide really focused on how the 2-year contract we signed for that cluster of B200s, the value of the contract really covered the face value of the GPU. So anyways, I provided the clarity on the CapEx there.

So you got another 1,500 GPs in the pipeline, 1,500 times -- sorry, that would be 3x the existing cluster size, so 3x $30 million, another $90 million, which again, you'd sort of finance. And that original cluster was with [ Dell ] through [ Dell], [ Dell Financial ] and they've been a great partner. So we may possibly finance through them or, again, typically 80% LTV. I hope that's really helpful, but that should give you the capital outlay to finance the GPU cloud business to get to that $200 million ARR figure.

Got it. And then I assume there's no capital outlay or capital expenditures on the Bitcoin mining side. And any plans to deemphasize that or possibly sell that off?

Correct. There's no CapEx allocated for Bitcoin mining. I mean, really 2025 was the year of bringing on that 18-plus exahash in Paraguay. And we had some credits with [ Bitmain ]. If you look at our presentations last year, we used this pledge. We pledged a lot of Bitcoin at $87,000 and when Bitcoin [indiscernible] was above that, we actually realized value from those credits.

And so with those credits, then we actually got some [ S21 SPs], which went to strategically upgrade some BUZZ miners. And so we're talking a few thousand units here. And really, just to get that breakeven as low as possible globally to be able to navigate new volatility, continue the cash flow. And that's really the only reason I had that price sensitivity slide in my section as well as is readily apparent if you're watching, if you like, wherever Bitcoin trades over the next quarter, you can always go back to my presentation, look at that slide and be like, okay, I should be doing about 50% margin after electrical cost, 45%, whatever the case may be, just to understand how that dual-engine strategy, how much cash is being thrown off.

But yes, no CapEx for the Bitcoin mining business. And I don't know what you mean by deemphasize. I mean it's, I think, a significant amount of cash flow generation we've got -- I know some of our peers have just because like [indiscernible] 23 exahash, whether they emphasize it or not, it's still there. So I assume The Street would like a bit of color and clarity that's still being well managed, and you're getting maximum value for all that capital that you've invested into that infrastructure to make sure that it is cash flowing as much as humanly possible.

And that's why I did take a couple of slides to point out that we still carefully -- any one of our deployments. We -- and I think it's worth noting, like zooming out we were once the world -- one of the world's largest are miners. We got into this cloud game GPU cloud game early in 2023 after the ether scaled and Bitcoin mining. So the very nature of orchestrating compute, it's -- to me, it's always been somewhat agnostic. I look at everything in dollars per watt hour. But yes, whatever we're going to be doing, we're going to be doing it well to ensure maximum upside for our shareholders. I hope that helps.

Next question, we'll go to the line of Mike Colonnese from H.C. Wainright.

Just a couple more on the GPU Cloud business for me. So as it relates to the data center infrastructure that will support the incremental chip deployments at the [ Dell ] data centers throughout Canada. Can you just give us an update as to where that infrastructure stands today? Is it ready to go? Are there other developments that need to happen at those sites to go ahead and support those additional GPU deployments?

Yes. So if you refer to the -- so the Winnipeg deployment, of course, is currently live. That's the first 500 GPUs. But moreover, when we put out a press release, I believe it was in early April for -- with Bell for the expansion to British Columbia on the West Coast. That was for the site and merit that side that the excitement around that was that the site was ready and effective for us April 1.

So the merit site is live, and so it's really an exercise of standing up the GPU clusters in there and going live with those. And again, just kind of circling back towards Joe's question is, we're going to kind of towards the final strokes of financing and announcing definitive for those cluster deals. And so yes, so you've got the capacity for the first cluster ready today at the Merit site and that the capacity for the second cluster in the [ Merit ] site will come online later this year, and you kind of see the time line that we've laid out in that. I would -- I'd refer you back to that chart -- slide where you kind of have the growth of the GPU clusters by type of [ GP ] by size GPU and by what quarter they're expected to be deployed. Yes, I hope that answers the question, Mike. Please let me know if there's any more clarity.

Great. So it sounds like the cadence of infrastructure development aligns well with pretty low risk to the schedule you laid out in your slide deck. Is that so to say?

Yes, yes, exactly. That's -- and that was an overtue of having the CapEx like quick time to market partnership with Bell. As you very well know, we are data center builders and operators ourselves, but having that at that logo, having that capacity having that Tier 3 bad capacity ready in a very near term allowed us to scale the cloud business as we saw outsized demand in the very near term and do so in a CapEx-light manner while we undertake our more longer tail CapEx-intensive either conversions of, for example, New Brunswick and [indiscernible] and little Toronto and, of course, now with the Gigafactory. So that was the whole strategy is to leverage that quicker time to market through Bell.

Makes perfect sense. And just one more for me. How should we think about gross margins for the GPU cloud business for those chips that will be deployed at the metal data centers? I know you mentioned you locked in a pretty favorable colocation fee you're paying [indiscernible], but just curious how we should think about the gross margins there.

Yes. I would put EBITDA north of 75%.

Time for a few more questions here. We'll go to the line of Fedor from B. Riley.

First of all, thank you very much for a very detailed presentation. Most of my questions already upfront answered, but I do have one on 320 megawatts newly acquired land that supports Gigafactory expansion.

Given the recent announcement, my question is, when did you start conversation or for ranging with capacity to potential tenant? And who you -- what kind of tenant you're looking at? And if you can just outline demand from Canadian potential tenants or maybe from U.S.? And yes, who are you reaching out and how far you are in the process?

I think I alluded to the -- thanks for the question. I think I alluded to the Gigafactory say really the crown jewel and the Canadian sovereign strategy. So there's -- given the proximity of the site, the GTA, Greater Toronto area is what GTA refers to is really like the prime area in Canada, this corridor really it's in the realm of the [ Vector ] Institute. There is -- and the short answer to your question is stay tuned.

Really, the inaugural announcement was to share with The Street, the capacity, the power and the land. And so specificity on who are the types of climb, so you'll have to stay tuned. But I can say directionally that this would be a phenomenal site even for having some government tenants in there as well.

And of course, in our partnership with Bell Canada, there's a -- although this is a HIVE and BUZZ-owned site, which is distinct from the Bell AI fabric sites, we've been co-locating. This is our own site that we're developing and building. We've got a phenomenal partnership with Bell. They have many Canadian enterprises that would love to have residency at the site. So I know you want more color, but it's one of those questions you just going to have to stay tuned. And if there's anything else that you'd like to ask, please let me know.

All right. Time for one final quick question. We'll go to the line of Stephen Glagola from KBW.

I just wanted to circle back also here on the 320-megawatt your Toronto site. One, can you maybe clarify what full allocation means for the 320 megawatts gross capacity? And then second, you mentioned graded load studies in the prior Q&A. Could you maybe provide more details across interconnection and permitting, including like where you stand in the ISO process and any remaining permits across zoning, building environmental required to begin construction or support customer contracts?

Yes. So quickly, Stephen, we're -- this isn't like the guys at semi analysis, they like to go ask all those questions. And there's just a certain level of disclosure that we're going to provide and then there is a level of disclosure that's sort of beyond the realm of what we're going to provide at this time.

So it's really stay tuned. I mean we've put a very thoughtfully curated press release. I spoke to it in my presentation, and that's the level of disclosure that's been shared with The Street. You e-mail the same questions to me a few weeks ago and I think my answer was much along the same lines. So I'm sure you'd love to come to the site and take a bunch of pictures too.

But you know what, that's not going to happen yet. So you'll get an invite when we do public walk-throughs, et cetera. But in the meantime, you're just going to have to hang tight body. So I think it's also covered in my sight -- sorry, it's covered in my slide. You put 240 megawatts of IT load if you assume 1.3 POE, and you could work out your revenue figures based on -- I mean it is a primary market.

So you could use your dollar per kilowatt, you go but $150 in there, if you want to get to revenue projections. I'm not sure if that was your question, but I hope that's helpful. And I want to be helpful, but sometimes people will ask questions that are just far beyond the scope of disclosure at any moment in time. So I just -- with all respect, I have to call that out.

Thanks, Steve. And that concludes our Q&A session. in our fiscal Q4 and full year 2026 earnings call. Thank you to our analysts, all of our attendees for joining. We look forward to speaking to you again soon.

Hello, and welcome to today's webcast on HIVE Digital Technologies financial results for the quarter ended December 31, 2025. My name is Nathan Fast, Director of Marketing and Branding at HIVE. I'll be your moderator for today's call.

Before we get started on Slide 2, I would like to briefly note the disclosures for today's presentation. Except for statements of historical facts, this presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. Words such as expects, believes and similar expressions identify these statements. Actual results could differ materially, and we disclaim any obligation to update them, except as required by law.

For a full discussion of risk factors, please refer to our most recent SEC filings at sec.gov. In addition to discussing results that are calculated in accordance with GAAP, we will also reference certain non-GAAP financial measures, including adjusted EBITDA, adjusted net income and free cash flow. Management uses these metrics to evaluate operating performance and believes they provide investors with additional insight and they're presented for supplemental purposes only and should not be considered in isolation from GAAP results. Reconciliations to the nearest GAAP measures are included in the appendix to this presentation and in the press release and Form 8-K furnished to the SEC.

I would now like to hand the presentation over to Mr. Frank Holmes for a macro recap of the quarter. Frank?

Thank you, Nathan, and welcome to the show by HIVE. I've got lots to share with you as the Co-Founder and Executive Chairman and give you sort of an overview of what I'm seeing in the geopolitics around the world and some of the glitches that have happened that have really impacted the crypto ecosystem. And I'll give you also somewhat we'd like to call the standard deviations that this is a great buying opportunity based just on the math of markets that when things become overextended on the downside or upside.

But before getting into this detail, I want to share with you that volatility is incredible. And it's also predictable when you take a look at it and you update this every quarter. But the S&P daily is 1% of the time. That means 70% of the time, it's a nonevent for the S&P to go up and down 1%. And over 10 days, 3%.

Gold is the same as the S&P 500, but Bitcoin is 3x greater. And that's something just to recognize that the stocks that are in technology, a new innovation like Bitcoin, it just will experience greater volatility. Tesla used to have this 21% volatility over 10 days until it became part of the S&P 500. But you can still see that on a daily basis, it's 4x greater than the volatility of the S&P 500 or gold bullion. And we can also see here that over 10 days, it's 11%. 11% is a normal volatility.

Strategy, what is interesting for me is because strategy was always higher volatility than HIVE. And now it's a daily basis is plus or minus 5%. But over 10 days, HIVE has a greater volatility, which surprises me because we do not have the same balance sheet. We have a very conservative balance sheet and a very conservative pragmatic way of managing the company's growth.

So here's the team. They're all working hard. Aydin Kilic, Darcy, our CEO and CFO, the longest-standing, I think, CEO and CFO in the industry. So I'm very happy about that. And Nathan has joined us a little over a year ago.

And now let's get into this macro recap. Here's the picture. And this picture is really important in respect that we grew our business from 6 exahash to 25 exahash in 2025. And we are 2% of the global network. And you can see that one of the investment bankers there, Jamie Brown, was on the ground floor of the creation of the idea of HIVE. And then you can see some of our directors are there and other key employees, our Chief Operating Officer and the President of Paraguay's operations, and you can see Nico's in there.

And you can see Aydin Kilic, our CEO; and Johanna Thornblad, who's the President of Operations in Sweden. So we're really proud and happy that we've been able to grow this business, and we still have more growth ahead in 2026, very pragmatic and thoughtful expansion. And behind there is a 5 kilometer, but a total length of 7 kilometers, 5-mile dam. It's the largest dam in the Western Hemisphere. It's able to generate about 14 gigawatts of electricity, half to Brazil, half to Paraguay.

So we recently came out with a production in January, and that was -- Bitcoin produces 191% year-over-year growth. Even though the difficulty went up, we've been able to maintain this substantial growth. And it's helped us with Bitcoin coming off over the past 4 months because we have this economies of scale. We've also been able to drive down our fleet efficiency. And when you see Joules per Terahash, that really is the amount of energy that you have to pay for that's running those machines and how the ASIC machines, how fuel efficient.

And it's really remarkable, if you went back 10 years ago, it was about 300 joules, and now we're down with like Moore's Law, it's only 17.5 joules. And the next generation is going to take this down to 11. So it's able to manage if you have new chips, you're able to manage the halving that takes place in the Bitcoin ecosystem.

And so we're more than 2%, and we're happy with that. Steve Jobs had this wonderful speech. And in that speech at Stanford, he said, you cannot -- you can't connect the dots looking forward. You can only connect the dots looking backwards. So if you have to trust that the dots will somehow connect in your future, it's important that you look back. So let's look back at some of the dots that what's happened to the Bitcoin ecosystem in the past 4 months.

On October 10, called 10/10, Bitcoin knock-out. This is a total crypto market cap fell $350 billion because it's alleged that is like Tyson punch, but the Binance had a faulty algo that basically triggered wrong accounts, whatever, and they blew out $19 billion worth of Bitcoin. And this knocked off and created contagion. What I didn't see it in front page of publications. I heard about it. I thought it was more noise, but actually, it's come out to be quite significant.

And so I'm going to show you how by connecting the dots. So October 10, and our CEO, Aydin Kilic has been at the Bitcoin Conference in Hong Kong. And this made the center stage, and there was lots of conversations and interviews regarding it. It's been dismissed by Binance and actually, they would. But what I want to explain to you is that there was this flash crash. It was only $19 billion. No, it's more than 10x FTX blow up. It's really quite significant.

And that triggers margin calls in North America. And then we have ETFs. The ETFs get amplified because almost $200 billion of retail institutional money rolled into a suite of different asset managers with their Bitcoin ETF. And they've been really hurt with a 50% decline. They just become -- they lose this thing called trust. And really the find out that supposedly a hedge fund made $1.2 billion on this flash crash. And other funds that weren't supposed to get hurt, they got hurt, but Binance made them hold with [ $300 million ] injection, there's no big positive amplification to all the ETFs that felt this stress.

So this was a big conversation and Binance comes out at the end of January and discusses it. But what's really important is that we've seen this before, and I know I've seen this in the futures market back in the early '80s. We've seen it in the gold markets where major banks would turn around and spoof.

And -- but you would have a court system and a court system that go after those traders and then they got charged and they were found guilty and then fines were paid by firms like JPMorgan. There was a -- that builds trust. So it's just -- there has to be a mechanism to -- if you have a rogue, they call this an algo. But I always ask why did that algo didn't work on April 2? Oh, it's all because of Trump and Trump went anti-China during that time period.

I think that's just too easy to go blame Trump because on April 2, those algos didn't cause this crisis that the magnitude that took place on October 10, 10/10. So that's what's happened. So I went back and looked at what happened in 1 hour, what -- how fast it just turned around and hit and it only grew that day.

So in Consensus Hong Kong this past week, Binance's Richard Teng breaks down the 10/10 nightmare that rocked crypto, but really doesn't explain to what hedge funds and a lot of the chatter that was there that our CEO was listening to and giving us more color on it. It never made the front page of the Wall Street Journal, but Sam Bankman-Fried did. And this is a factor of 10x greater. So it's really disappointing, but it is what it is.

And the CEO, he's worth $88 billion according to Forbes. And other people say he's worth $30 billion to $40 billion, $50 billion. He's got this brilliant mind, what he's been able to do. But it doesn't make sense that this is an unregulated combination where the exchange and the investment broker is combined. You can't really do that. The New York Stock Exchange doesn't own money management.

It's like merging the New York Stock Exchange with BlackRock and all of a sudden, then BlackRock has access to what the trades are and the traders know what the fund flows are doing. That's the difficulty in this unregulated Binance exchange, which say they trade trillions of dollars in notional value in the Bitcoin ecosystem.

So I hope that there's a better clarity on what takes place. But it's 4 months later, and it's interesting that they come with explaining at the end of January -- and usually, when these crises happen, it's about 4 months later that we get a bottom in the ecosystem. But what happened after this, it's interesting to me in connecting the dots is Jim Chanos comes out and short the Bitcoin miners, short the NVIDIA, short the HPC, the hyperscalers, there's too much debt and Michael Burry is coming out. He came out a couple of weeks ago again.

He short this market. And it really starts to grow that -- this negativity on the ecosystem. And you can see all these headlines. Chanos warns of AI pullback, absurd Bitcoin treasury companies. Michael Burry's latest argument. Chanos is going after Michael Saylor, shorting them, Michael Burry, same thing and all these legendary people.

And to me, I just looked at -- I remember commenting that there's something weird that there's just so many people that are all of a sudden negative on the AI, but I know that the demand is so big and the ability to build out is going to take more time. So I don't see this, the level of this negativity, but the Binance breakdown of the ecosystem that flash crash was part of this.

And I always love this scene. It's from the Superman movie at the time. And it's basically saying where they have these monkeys that are out there making all the chats on Instagram and YouTube and X and Super -- this is about Superman's credibilities being destroyed. Well, the same thing happened out of nowhere all this negativity was showing up on Instagram and YouTube and X. And so I just -- it was a short-term fuse, but you look for as a money manager, is this sustainable? Is it real? And it became the trade.

Well, let's look and connect the dots. We go back to October 10, and you can see starting at the beginning of October, the stocks have this big rise and CoreWeave goes through the roof, and we see HIVE is on a tear rising up and Marathon has the bounce and we see Core Scientific is rallying. But then after October 10 starts this Binance algorithm, it's like COVID contagion to tech stocks.

And you can see what took place with this, this domino effect, and this impacted Bitcoin prices coming off. And I still see every day on CNBC regarding the negativity of the hyperscalers and the Bitcoin treasury companies. But there's these inflection points that happen. They last about 4 months. And then we have another cloud that happens at the same time.

The U.S. Senate Committee delays the crypto bill after opposition from Coinbase, Brian Armstrong. Why? It's because a lot of the banking system in America is not really cognizant of China's war against the U.S. dollar. They are just not aware that China has taken the BRICS nations and weaponized that trade is not in U.S. dollars, it's in the yuan and don't own one enforcing and pushing that these central banks basically sell their U.S. dollars. And now we're seeing gold become the biggest foreign exchange in many of these countries, central banks.

And we're seeing now the thought process of offering a reward mechanism or would it be like a money market fund for stablecoins so that the U.S. government would be able to have their stablecoin and places like Coinbase, they'd be able to pay, they call it rewards. But really, to me, in the money management business, it's like a money market fund. But it's very significant for the growth of mutual funds and then ETFs. This would be very significant for the crypto ecosystem and also for the U.S. dollar because we've seen the success of Tether.

Tether is phenomenal that bad country, bad policies like Venezuela, like Argentina, the Lebanon, you can see that Turkey, the currencies are being devalued and so people turn around and they bought Tether stablecoin and they're protected with U.S. dollars.

Now they've been -- the Tether gold coin basically has been growing at a phenomenal rate. So we can see hundreds of billions of dollars going into those -- the success of what Tether has. But Tether doesn't pay a coupon, and therefore, it's not a money market fund, and therefore, it's not a security. So it's able to become a dominant like a U.S. dollar currency that people can digitally move money all around the world without the big banks turning around and delaying the payments and saying their AML/KYC concerns, et cetera, which is rightfully so, but it's becoming just so onerous to move money around, especially between countries and repatriating.

So if I am a worker here from Mexico and my family needs money back in Mexico, I can do it much faster with a stablecoin. I can -- much faster than any other way and the repatriation of that money helps these other families in other countries. Well, along comes the stablecoin and Coinbase wants to pay rewards. They want to basically make it like it's a money market fund.

The banks don't want that because they want their stablecoin to get big. And before they turn around to allow a coupon and they keep saying, what will cause a crash, people will leave the banks and they'll go to the stablecoins because they're paying a coupon. It will hurt banks. You know what, I listened to this, and it happened a long time ago when banks weren't able to pay the coupon in 1980 as interest rates soared to 20%, but money market funds were.

And money market funds grew dramatically, which only helped the growth of mutual funds and equity funds. So that wasn't a big loss to banks. But what did happen is the S&Ls, they were allowed to pay a higher yield and the banks didn't like that, and they grew. And they had an S&L crisis. So I think it wasn't a banking crisis so much as the banks do not want competition. They do not want this fintech. And really, Bitcoin is a spoke in the wheel of fintech.

Coinbase is a critical spoke in that wheel of building fintech around the world is the way, in my opinion, as a money manager looking at what's going on. We're just going through this process.

I think that paying rewards, it will win, it will get through, and this will be the reprieve. But it's a battle between the self-interest of the banking industry and lobbying groups and fintech growing. So we have this backdrop. We have the carry trade unwinding, which is about $500 billion throughout the month of December, especially January. So that's been behind us now.

We have the Binance igniting a huge meltdown in the Bitcoin ecosystem, breaking that trust factor. And then we have this last Coinbase. Well, let me share with you, we are down 2 standard deviations. And it's only happened a couple of times. You can see when China did its attack on the crypto ecosystem, and we had Bitcoin fall and the miners fall and then America end up benefiting becoming the biggest Bitcoin miners in the world. Then we had the Celsius, the FTX blown up, and we saw the prices fall 1 and then 2 standard deviations.

Now we're down to 1.64 standard deviations. And that says to me that mathematically, what is 1 standard deviation over a 20 rolling day period, it's about 17%. So it's suggesting here that we could get from here a rise of 30% to 40%. We could get a higher rise if the act gets passed and with -- with where they can pay rewards on stablecoins, that would be a big boom, and we would probably see this go up 2 standard deviations.

So you can see that in a bullish cycle, it goes up 2 standard deviations more than it falls. But we are at an inflection point of accumulation is the smartest option, not to capitulate and sell out. So then I asked, let's look at HIVE. Same thing. You can see that HIVE was up at the very top here going into the beginning of October. We had a big run because we went from 6 exahash to 25 exahash and the world loved it, and we're going through this rerating until the Binance faulty bot they have blew up their system.

And then we fell. And we had a rally and now we're down once again over 1 standard deviation. And I think that we're an attractive buy from based on just the math of markets. Something else that is really important. We have never leveraged our balance sheet with incredible debt to go buy Bitcoin or to go and do contracts with for high-performance computing. We have not done that because we're conservative because we know the volatility.

We know that there's building out Tier 3 data centers is fraught with construction difficulty. You have to be very pragmatic and thoughtful. And so we've not done this huge debt financing on our balance sheet. But what we have done is what we make the press release today is that we are -- HIVE isn't chasing AI fairy tales. It's building towards $140 million annualized GPU cloud revenue from measured steps. Today's $30 million 2-year contract secures the initial 504 GPU Manitoba rollout in partnership with Bell Canada, lifting HPC ARR.

And the reason why I share this with you because a year ago, we had a run rate of about $1 million a month, then we got it up to pushing $2 million a month, and this is going to take us to $3 million a month, and our long-term vision is $2 million -- $10 million a month. And we're doing it in a measured way. This is a Tier 1 data center we bought. It's going through the transformation to Tier 3, and we get people knocking on our door that we know that if we had it up and running today, we have contracts for 5 years to buy and that give us big contracts. So it's interesting.

Our strategy is just different than other people that are going out, getting a hyperscaler, giving you about a $0.14 -- when we look at the math of this, a 10-year contract, a 15-year contract, but we're trying to get as much of the upside besides a tier HPC colocation that we know we've been able to build with high margin than people around the world. We've built the 10,000 customers in 80 countries are using our chips and mining by the hours.

Some give us contracts for longer time periods. What we're seeing now is that once this is built, this will be solved. That I mean, we'll sell the asset, but the demand for our GPU chips at much higher prices than where Bitcoin revenue is per hour will be done. So we feel very excited about it and straightforward of how our vision is just different than other people. And that's where we're staying focused. This is another build-out. This is the data centers that we have.

And as you can see, we have them in New Brunswick, and this will go through a conversion, we bought more land, and this will go from 70 megawatts. What's interesting is it will be about 50 megawatts, we'll actually be able to do the HPC because a lot of people don't realize that when you go from Bitcoin mining to Tier 3 HPC or Tier 4, the bulk of your energy is used for air conditioning, 40%, not 5%, but you're now 40% of the electrical bill because those NVIDIA chips consume a lot of power.

They give up tremendous amounts of heat. You have to be up 24/7. It's a very different business model. But we're plotting along, and we feel very good about -- and our President is seeing nothing but big demand coming in. If we were up and running today, everything will be taken for our chips at very attractive contract prices. So it's now about being pragmatic.

You get your chillers, you start -- you have to preorder because the transformers -- there's a backlog for transformers. There's a backlog for the equipment you need for building substations today. There's a backlog for the special server racks and what you find now with HPC with the NVIDIA chips, the server racks are heavier. So now you have to build cement floors that are thicker. And so it's not easy to say, okay, we'll just convert. No, it's very thoughtful, and that's what we're doing as we're managing our cash.

This is to share with you that data centers are continuing to grow, and they're a very big part of the GDP growth in America and the GDP growth in Canada. And what people have to be listening to is that it's not just Open Chat that's looking for these data centers. It's also the military spend.

And the NATO countries have now gone to 5% of their GDP. Well, a lot of the new weaponry needs data centers that are high-performance computing. And if it's military, they need Tier 4, which is another level of security and backup. So we see globally the demand for these data centers is not just these wonderful new platforms like Perplexity or Claude and Grok, I love Grok, I love Claude and Open Chat, it's military spending.

And then you have countries are saying, we only want our data centers and data in our country, we want it sovereign. So this creates another pent-up demand. And so we think we're in the sweet spot of being the biggest player in Canada at this stage, and we'll grow this and move this and we'll become the biggest player in Paraguay is our vision.

This is just to give you some color about the hyperscalers are ramping up their CapEx. And this gets all this negative news. And I think that who's making this negative news is just really helpful for a short-term trade of being short. But the Metas and the Microsoft, they've not been spending like, well, I should have had here is Oracle, too, and CoreWeave, that spending will continue. And I see the reason for it is the backup demand is just immense.

So here is to give you an idea of the future shock, the scale and speed of AI disruption, $100 billion hyperscalers are pouring into the AI infrastructure. That's just in America. You got to think about the rest of the world. $25 billion market impact, revenue shift from NVIDIA dominance to Chinese domestic chipmakers. These are all big real issues in the global race.

What China has done for the past 10 years has exploded in sources of energy. They have been building hydro dams. Spain has been unwinding hydro dams, 2,000 of them relying only on solar and wind, and that's created their own energy crisis. But you're not seeing that in China who continues to build from hydro and dams of rerouting glacier water from the Himalayas down the rivers to basically create these massive dams and this hydro is so key, so now they can ramp up their data center business.

Here in America, we've got more HPC data centers, but we've got to ramp up both sources of energy and be innovative and creative with that. And now it's nuclear energy is cool. Now nuclear energy is not the bad word. So things are changing, but the idea, it's unprecedented. And every year, I spend a week at Harvard, with 180 CEOs from 80 countries doing cases, and it's interesting to see that AI dominated all the cases. Leadership disruption, what Microsoft had to do in Europe. Greece is now trying to do a huge educational push.

Open Chat is partnering with them. The Onassis Foundation is partnering. A former [ McKinsey ] consultant went to Harvard. They're doing everything to fast track the kids' education so that they can participate in this growth in Open Chat and anything to do with AI. This is recognizing the future demand for accelerated computing and graphics processing and NVIDIA began designing GPUs specifically tailored for to meet these needs.

You know their big move was their pivot was 2010. And at Harvard, one of the cases was on Jensen. And what I did know is that Jensen's parents sent him to a private school in America, and they didn't realize it was a reform school, a Baptist reform school. So that made him really get tough and resilient, and that's what the whole idea of NVIDIA. But they did this big pivot in 2010. 26 -- we're talking about, what, 16 years ago, and then AMD is now related, but the CEO of AMD was a part of another case, and she is related to Jensen.

And she has a PhD electrical engineer from MIT. And part of their pivot and AMD is to go in this space, but they're still very far behind where NVIDIA is. So I think we really are in a secular bull market on the adoption and the build-out necessary for AI and stay -- look to buy the dips.

I want to thank these investors. Two Sigma Investments, there's a quant shop and Chicago Park Employees and Tidal Investments, Citadel Advisors, Schwab Corporation, that's individual investors. And it just amazes me that Schwab gets sold down and all the fintech just recently, this negative narrative because fintech is going to be disruptive with AI for all their client business. I don't think so.

I think if anything, AI is going to probably help on the overall compliance and the complexity of compliance and monitoring and things like that. And we're seeing KPMG has to be honored by independent auditors. And KPMG is going after their auditors for not getting lower audit bill for -- but the KPMG is ahead in using AI.

So to say that AI is a bubble and it's all over, it's just market chitter-chatter for trading just short. And so I remain very, very bullish. There's Aydin with the Chief Minister in Paraguay, very important in the overall business development to be very close. We regularly go meet with ministers in Paraguay.

Our President, Gabriel Lamas, and I met with the ambassador from Paraguay, who's based in Washington, D.C., what their vision is. And they have a big vision of making Paraguay the dominant AI infrastructure build-out for all of Latin America. They need other sources of electricity. They know that. They hope to attract solar farms and solar independent electrical grid.

They're looking at -- they changed totally, the cost of energy is dropping, long-term contracts, something they didn't give when we first went there, but now they are. We are building out Tier 1 data centers so that they have the runway for Tier 3.

During this process from Tier 1 to Tier 3, you need to get dark fiber built over the country, just like we know this has to happen in Eastern Canada. You can't move the data from Tier 3 data centers around the world unless you have dark fiber because these large language models have so much compression of data in them.

So, that's what we're doing. Now I'm going to turn over to Aydin Kilic to really give you an in-depth analysis of the company. And I hope that my presentation today is to give you some color about this incredible meltdown that's happened, what was the catalyst. We're probably mathematically at the bottom. And hopefully, going forward, I believe that we're going to trade much higher and HIVE is in a strong balance sheet position to monetize that growth with 10 exahash in Paraguay and huge upside in Canada and Sweden and the HPC business.

Aydin, take it away.

Frank, thank you very much for the insightful macro summary and now for an executive overview of this quarter. Now it was a really exciting quarter for us, and this is a photo from a recent visit to Paraguay. This is Minister of Foreign Affairs for Paraguay, Ruben Ramirez Lezcano, who you may have recognized in the recent Status of Forces Agreement signing between Paraguay and the U.S.A. with, of course, Secretary, Marco Rubio. More on that later.

Okay. So it was a record quarter for HIVE, $93 million of total revenue. Of that $32 million of gross operating margin. Now while we did have a $91 million net loss, that was mostly noncash charges, $57 million in depreciation. Of course, we brought on a lot of new hardware online in Paraguay as we scaled the 300 megawatts and also a $31 million noncash charge on change in fair value derivatives, mostly driven by changes in Bitcoin price. So on an adjusted EBITDA basis, $5.7 million and ending the quarter with 481 Bitcoin on the treasury. So again, record revenue for HIVE and really proud of the team.

Let's jump into the next slide. On an annualized basis, we realized $385 million ARR for the quarter, 879 Bitcoin mined. We realized 25 exahash of installed capacity, operated at an average of 22.8 exahash for the quarter as we had ramped up towards 25 exahash. And with the colder months, you have some temporary curtailments due to the very cold weather in the Canadian operations.

New Brunswick can occasionally and Lachute, but very happy Paraguay was performing with nearly 100% uptime. And of course, being in the Southern Hemisphere, when it's cold and there's cold snaps in the north in the Southern Hemisphere, it's actually summertime. So being geographically diversified has its benefits. Ladies and gentlemen, 440 megawatts of operating capacity with an additional 100-megawatt PPA, we announced the signing of that late last year. And long lead items such as transformers have been ordered, and we expect that to come online September of this year.

Now on the BUZZ side, another very solid quarter. Looks to me, $5 million revenue for the quarter, keeping, tracking the $20 million ARR, and we are on track to reach our target of 11,000 GPUs on the BUZZ cloud by the end of this year. Currently, 5,000 GPUs, will be adding 6,000 GPUs this year as well that target of $225 million ARR between the GPU cloud business and the HPC colo is on track as well. And we have some very exciting news.

Next slide, massive progress on that front, 70% increase to our HPC ARR. The $20 million ARR will be at $35 million ARR at the end of this quarter. And that comes from the signing of a 2-year contract for our incoming NVIDIA Blackwell B200 GPUs. So we announced in November that we ordered a 63 node cluster with NVIDIA Blackwell. They were destined for Manitoba, our first site with Bell. These GPUs are now fully contracted, receiving a deposit this week and the GPUs will go live this quarter in March and therefore, be cash flowing.

So we will be ending the current quarter period end March 31 with $35 million ARR, again, which is a 70% increase from the current quarter or reporting quarter of December 31. So huge news. Craig Tavares and the BUZZ team have done an absolutely phenomenal job. And I also want to point out that this is a very nimble, agile CapEx-light strategy allows to scale the GPU cloud business with the infrastructure that Bell AI Fabric is bringing online, and we've had very, very attractive single-digit lease-to-own financing on the GPUs themselves.

So no CapEx upfront for the GPUs. The entire full value of the GPUs, we are effectively leasing with a $1 buyout. So it works out like a finance, let's say when you finance a car with 0 down and single-digit interest. So very attractive. Again, Craig and the BUZZ team have done an absolutely tremendous job and more great news to come. Please stay tuned.

Let's hop into the next slide. We have a vertically integrated growth strategy. We have the land, the power, the data centers, whether it's ASICs or GPUs, we build, we operate and we optimize. So on the Bitcoin mining business this quarter, we realized $350 million ARR, mining approximately 10 Bitcoin a day in our Tier 1 data centers globally.

On the HPC business, as mentioned, our new benchmark is $35 million ARR in the current quarter, March 30 -- end March 31, and that will scale to $225 million. We're going to have a closer look at that very shortly, and that's a Tier 3 data center strategy.

Another nice Easter egg that we are providing the Street an update on is we realized $14 million of value from our Bitcoin pledge. You may recall, we had a substantial amount of Bitcoin, almost 1,400 Bitcoin pledged at $87,000. What that meant was we put up our Bitcoin at $87,000 to buy our ASICs, which was for expansion to 25 exahash in Paraguay. Once that Bitcoin was pledged at $87,000, that was it.

However, we had an option to buy back the Bitcoin at $87,000 when Bitcoin rallied above that price. And so we did that and realized $14 million of value, which is great news. And we call that our dynamic HODL strategy. We're going to provide a bit more color, but I just want to clarify, there's no cash call. There's no obligation. There's nothing like that. It's a free call option is what it was. Locked in the price at $87,000. Any upside beyond that, it was at our discretion, our call option to exercise, and we crystallized $14 million of value. So very exciting news.

Next slide, please. An overall footprint of the HPC various operating jurisdictions and data centers, you can kind of see how it ramps up to $225 million, which is a target for the end of this year between GPU cloud and HPC. The HPE conversion would be for New Brunswick to be converted to 50 megawatts of critical IT load as a Tier 3 data center, adding $85 million of ARR to $140 million coming from the GPU cloud spread out over the various Canadian facilities and showing how we increment from the current 5,000 GPUs to 11,000 GPUs. Again, we recently announced that 504 GPU contracts.

So let's go to the next slide. So here's what the growth of the HPC revenue looks like on a time series basis. We provided this projection last quarter as well. And as you can see here, for every 1,000 B200 GPU cluster, we'd be adding $20 million of ARR. Now keep in mind, the prevailing market rate at the time was about $2.20 per GPU per hour. And so this ramp from $20 million to $140 million ARR came along with 6,000 NVIDIA B200s being brought online. And then in addition to that, the $85 million estimated ARR from the conversion of New Brunswick to hyperscale and colo.

However, next slide, please. Due to the very strong market demand, the realized value of the GPU contract that the BUZZ team secured was 30% above forecasted prices. So what that means is where previously we projected $20 million ARR per 1,000 GPUs, we realized $15 million of ARR for 500 GPUs. So that is tremendous, 30% above forecasted. Again, this is liquid-cooled GPUs, and this works out very well. Here's an illustration of what that does for our projected revenue. Let's hop to the next slide.

I do want to say this is a potential and the team did a tremendous job. There's strong market demand right now. And so this is a blue-sky slide where if we were to scale the rest of the 6,000 GPUs at the same rate that the current deal was secured at, it would actually bring the GPU cloud ARR potentially up to $200 million by the end of this year. And then in addition to that, roughly $85 million from the NB colo, $285 million potential.

Now again, we're going to stick with our baseline projections on the previous slide, just tremendous job by the BUZZ team where they realized the 30% higher contract value due to -- driven -- strong market demand. But it's not just having strong market demand. Let's hop to the next slide.

Craig and the team have done a phenomenal job building out the BUZZ HPC cloud, which was awarded Bronze on ClusterMAX, which is in very good company with other very reputable clouds in the Bronze category. And you'll note a lot of peers actually were underperforming or even unavailable and some very well-known clouds in those categories that BUZZ outperformed. Next slide.

And I recently had a call with an analyst who didn't quite grasp what that meant. Well, when you're just renting GPUs bare metal, what that means is the user has to use an SSH key to get secure access into a GPU environment. They have to install the operating system. And really, it's just bare bones. And so not everybody -- if you're a model builder or a researcher, that's different. That's -- loading up an operating environment and clustering GPUs yourself is different than actually doing your LLM work. So what you want is this to be done for you. So you're getting a managed AI service.

You have Kubernetes, you have Slurm. These are 2 integral components to having a proper cloud. So it's very easy for you to have this elastic GPU resource, whether you want 1, 8, 32 GPUs when they're properly orchestrated, they all work as 1 elastic computing resource. And so that is virtue of having proper cloud technology, which the BUZZ team has done a remarkable job. And so that's how we're able to attract these great clients and have strong demand. So again, a phenomenal job by the BUZZ team. And really, I think there's going to be some more exciting announcements in the months to come. So stay tuned as we execute and march toward those revenue growth targets.

Next slide. So on the Bitcoin mining side of the business, for the recent month of January, we did about 9.6 Bitcoin a day, again, 440 megawatts globally. We lead the sector in low G&A per Bitcoin mined, maintaining optimized ROIC Bitcoin modeling -- Bitcoin mining model.

Let's go to the next slide. So as you know, we've got another 100 megawatt PPA that was announced in Paraguay. Paraguay has been very strongly aligned with the U.S. In December, Minister Ruben Ramirez Lezcano signed the Status of Forces Agreement with Marco Rubio, Secretary of State for the U.S. in a very momentous occasion. So it just shows the very strong alignment between Paraguay and the U.S. And Paraguay is really emerging as, I believe, one of the strongest U.S. allies, not only in Latin America but globally.

The SOFA, Status of Forces Agreement, is only held by a handful of countries globally with the U.S. And so really emphasizes Paraguay as a stable and safe jurisdiction for foreign investment. And we see a very bright future for Tier 1 and Tier 3 data centers in Paraguay. Stand by for some very exciting updates over the course of the next few months as well as we continue to have a very bullish outlook on our investment and expansion into Paraguay.

Next slide, please. This is a summary of, of course, our 440 megawatts of operating capacity worldwide and then the additional 100 megawatts we'll be bringing on. It's actually Phase 3 of our Yguazu site in Paraguay. Next slide.

Okay. Here, we're going to talk about that $14 million in realized value from our Bitcoin pledge. So as we previously discussed, we have pledged Bitcoin at numerous prices. And we had a large pledge of approximately 1,400 Bitcoin at $87,000. So what that meant was when we purchased our ASICs, we put up Bitcoin when it was $87,000, and we have the opportunity to buy it back at that same price. We redeemed our Bitcoin at $93,000, at $110,000 and $123,000 with respect to the pledge. And on that, we realized a value of approximately $14 million.

Then we took that realized value and translated it into approximately 3,800 Bitmain S21 XP air cooled, which then replaced our BUZZMiners, which very recently, as we've seen a contraction in hash price, those BUZZMiners have faithfully served us in hash for years and years. Approaching end of life, they've been upgraded. And so what it did is, it upgraded and increased our global fleet efficiency from 17.5 to 16.7 Joules per Terahash. What does that mean?

We used our pledge strategy to get a cashless realized value of $14 million, turn that into over 300 new generation ASICs and effectively lowered our global cost of mining in a bear market by 5% through dynamic HODL treasury management. And so it's just how we operate at HIVE.

We're, again, deeply analytical. We very much study hash price and have a dynamic HODL strategy that allows us to realize value beyond our mining, but also through treasury management. So I hope this is really helpful for the analysts.

And in addition to that, we still have 540 Bitcoin at the $87,000 strike price. Now with Bitcoin at about $66,000 as of time recording, that doesn't mean we have to put up any money. It's downside protection. We already did not pledge Bitcoin at that price. So in the current climate it's downside protection.

If Bitcoin happens to rally in the next couple of months, beyond $87,000, we can realize further value. So really happy with how this all played out. Again, we've been through numerous bear markets. HIVE has been through Bitcoin halvings and Ethereum merge. We built our own ASIC miner with Intel. You learn a lot along the way when you've been through it all.

So let's hop into the next slide. Of course, mining economics have contracted a bit. We had the calamity from 10/10. The structural errors where the collateral coins held by Binance were effectively shorted and that led to auto deleveraging in October 10 and Bitcoin dropped from $126,000 to $105,000. But moreover, all those auto deleverage positions, a lot of people got washed out. Binance put up $300 million to make some people whole. A lot of retail investors took a hit though.

And recently, Binance put $1 billion to help support Bitcoin at the $60,000. So really, it was worse than the FTX crash. And it's just for people to be aware of why did Bitcoin sell off. And again, there's obviously broader market headwinds where we've seen a risk-off environment. And so as a result, we've updated the annualized mining margin analysis for all the shareholders and anyone watching this podcast, so, $30, $35 and $40 hash price, here are your projections. So current difficulty of $126 trillion with Bitcoin at $60,000 the left column, you've got a $30 hash price, well below where it is today. Today, Bitcoin is about $35 hash price. And if -- and then with Bitcoin at $80,000, we be a $40 hash price. So let's just start on the left column.

Even in a more bearish scenario of $30 hash price -- by the way, we did see hash price flip down to $27 in last week when Bitcoin hit down to $63,000. Keep in mind, difficulty was still $141 back then. We saw flash crash at $27 hash price. And so anyway, I just want to give context where has hash price been. Has it been as low as $30? Yes, for a moment in time. So nevertheless, we project that even at $30 hash price, we still have an annualized mining margin after direct operating cost of about $90 million. So still healthy margins. At $35 hash price, mining margin is $135 million and at $40 hash price, $180 million. So this helps you have an outlook of what it could look like in a contracted Bitcoin environment.

Like -- I saw Richard Teng, the CEO of Binance, actually speak at Consensus Hong Kong today actually. His accounting of it, he was quite stoic and he mentioned that you have these near-term measured in months and these calamities that happen in crypto. But when you look at the years horizon, the asset class consolidates and has grown in value. So it is another headwind that we will navigate, again, having low G&A, a very best-in-class mining operation amongst our global sites and great fleet efficiency. Again, that upgrade basis was done on a cashless basis, $14 million in realized value. I'm very proud of the HIVE team for all of the great scaling and very judicious and, in my opinion, expert-level Bitcoin mining. Next slide.

Again, this is a really helpful visual, just sort of like a math textbook. What is the fundamentals of Bitcoin mining? A lot of people understand it, but do they truly understand it? And so really, what you're trying to do is ROI in the first year to 1.5 years, and that's shown in the blue section. Your hash price does eventually commodify as more hashrate comes online, and there is an implicit breakeven and therefore, end of life cycle.

So your power cost, as that goes up and down, the higher your power cost, the shorter your x-axis, your horizon of useful economic life, the lower your power costs, the longer you can mine, therefore, the longer you can free cash flow. So anyways, it's just something to be aware of how does crypto mining work by design. Yes, you do upgrade your machines every 3 to 4 years, but we run them for as long as possible. Keep in mind, our BUZZMiners, those came online in 2022. So all those BUZZMiners have been mining for almost 4 years now. Let's hop into the next slide.

Again, a big part of our ROIC-driven ethos, also having low G&A. And so let's go to the next slide. Not all of our peers have reported yet, but just based on those that have, again, lowest G&A in the sector. By the way, I do want to point out, if you compare on a year-over-year basis, our G&A is up about 80%. However, our revenue is up over 300%, it is over tripled. And our corporate margin is about $30 million this quarter, is up about 40x from a year ago when the corporate margin was $700,000. So the point is even as we've scaled the business dramatically, our G&A has not grown nearly as much.

And so, again, we maintain that lean and mean mindset. And by the way, we have a Bitcoin mining business and an HPC business. So very proud of the entire executive team. We've had to huddle over 9 time zones every single day. We're in 2 hemispheres. We're in multiple continents. Let's go to the next slide.

Also, best value. If you look at our peers, where they're trading on a EV to exahash, it would place HIVE with a $3 billion multiple, everybody is going to say, yes, but everybody else has HPC and landing power. So do we refer back to the tremendous growth that we're experiencing and that we further have projected for the rest of the year on the HPC business and stable and steady cash flows even in the sort of temporary bearish Bitcoin mining climate that we see. And of course, Bitcoin is a very cyclical asset class. You really make hay when the sun shines. And so we'll be ready for the next bull run when it comes. But in the meantime, we'll be cash flowing.

Next slide. Darcy, longest-standing CFO in crypto mining. Over to you. Thank you.

Thank you, Aydin, and good morning, everyone, and thank you for joining us today. I will be walking you through the highlights of the quarter. We are providing certain non-GAAP measures in our presentation today. The company believes that these measures, while not a substitute for measures of performance prepared in accordance with U.S. GAAP, do provide investors with an improved ability to evaluate the underlying performance of the company. These measures do not have any standardized meaning prescribed under U.S. GAAP and therefore, may not be comparable to other issuers. Further details are found in the management discussion and analysis for the 3 and 6 months ended December 31, 2025.

Starting on the next slide, HIVE ended the December 31, 2025 quarter with 243.1 million shares, 2.6 million options, 13.6 million RSUs and 3 million warrants outstanding. I'll now walk through our financial results for the quarter ended December 31, 2025, beginning with key operational and financial metrics.

Q3 represented a quarter where we continued to execute operationally while navigating market volatility in digital assets. Our focus remains consistent, disciplined capital allocation, operational efficiency and cash-oriented returns on invested capital. Let's start with the headline financial outcomes on the next page.

For Q3, we generated $93.1 million in revenue with approximately 95% coming from hashrate services on our Bitcoin side and nearly $5 million contributed by HPC operations, demonstrating the scale we've achieved as we continue ramping toward higher hashrate and HPC expansion. Adjusted EBITDA remained positive at roughly $6 million, reinforcing that our operating model generates cash despite cyclical pricing conditions.

Operational output remains strong with approximately 884 Bitcoin equivalent produced, which is up from 719 in the prior quarter, supported by stable operations, strong uptime across our sites and the execution of our Paraguayan expansion. At quarter end, we held 481 Bitcoin on the balance sheet, reflecting our hybrid strategy of liquidity management and strategic digital asset exposure. These numbers reflect disciplined cost management, a focus on efficiency and the benefit of our diverse global footprint.

Now let's, on the next slide, take a look at how this operational performance translates into our balance sheet. HIVE takes pride in maintaining a healthy balance sheet. Turning to liquidity. We closed the quarter with approximately $14 million in cash and $14 million in digital currencies, bringing total current assets to about $91 million. Current liabilities stood at approximately $52 million, providing us with a healthy working capital position. This balance sheet supports our dual growth strategy, expansion in Paraguay and scaling our subsidiary, BUZZ HPC, while maintaining financial flexibility.

Our strategy remains conservative on leverage and disciplined on capital deployment. With that context, let's look at how our earnings metrics have evolved starting on the next slide. Shifting our focus to our gross operating margin on a year-over-year basis, comparing the results of this quarter to Q3 last year, our gross operating margin, which is calculated as total revenues minus direct operating and maintenance costs and HPC service fees, increased to $32.1 million in the most recent quarter compared to $5.3 million in Q3 last year. In this most recently completed quarter, we are reporting a basic loss of $0.38 per share compared to a net income of $0.53 per share reported for Q3 last year.

This reduction in earnings per share is largely driven by noncash accounting impacts such as the accelerated ASIC depreciation tied to our expansion in Paraguay, unrealized losses on investments and digital currencies held on the balance sheet and changes in the fair value of derivatives.

Taking a look at our revenue increases year-over-year on the next slide, we generated total revenue in the third quarter of fiscal 2026 of $93.1 million versus $29.2 million in the previous year's third quarter. On a year-over-year basis, revenue growth was supported by higher production scale and operational uptime.

Year-over-year, we saw a significant improvement in gross operating margin, expanding from roughly 18% to about 35%. This reflects the benefit of our efficiency initiatives, though it continues to move with Bitcoin pricing and network difficulty. It is important for investors to understand that our margin profile is heavily influenced by external variables, whether this be hash price, power costs and market volatility, while internally, we continue to focus on controllable drivers like uptime, fleet efficiency and SG&A discipline. Even in volatile market conditions, our goal is to maintain a structurally stronger operating model.

We're focused on expanding the structural margin, not chasing cyclical upside. And if we zoom in just the last 2 quarters, you'll see our continued strength on the next slide. Comparing our current fiscal Q3 quarter to the previous Q2 quarter, we generated revenue in fiscal 2026 Q3 of $93.1 million versus $87.3 million in the previous quarter. A slight increase in revenues versus the prior quarter was impacted by continued increases in exahash capacity from Paraguay in spite of digital asset price movements and changes in network difficulty.

Our gross operating margin decreased to $32.1 million or 35% in the most recent quarter compared to $42.4 million or 49% in the prior quarter's comparative. These quarter-over-quarter comparisons show margin compression relative to Q2, primarily reflecting digital asset price movements and timing effects rather than structural changes in our business.

Operationally, our facilities continued to perform well with strong uptime and efficiency metrics. What you're seeing here is market sensitivity. This is economics of the cycle, not a change in the trajectory of the business. As we scale toward higher hashrate and benefit from ongoing efficiency upgrades, we expect operating leverage to improve over time.

And on the next slide, I'd like to remind our stakeholders that our net income is comprised of our operational earnings or cash flow, plus our investment earnings, which includes realized and unrealized earnings, which often includes noncash charges. Our adjusted EBITDA for this quarter ended December 31, 2025, was $5.7 million compared with adjusted EBITDA of $82.9 million for the December 31, 2024 period. The largest contributor to the high adjusted EBITDA in the prior year was a $77.4 million unrealized gain on digital currencies. I will highlight again that adjusted EBITDA is a non-GAAP figure.

For this completed quarter, we experienced a loss of $91.3 million compared to a net income of $68.2 million in the previous year comparative. On earnings, year-over-year comparisons include significant noncash impacts. Specifically, we have accelerated ASIC depreciation tied to the Paraguayan expansion, which reduces accounting earnings in the near term. This accounting treatment aligns depreciation with asset utilization, but does not materially impact cash generation. Adjusted EBITDA, therefore, often provides a clearer representation of underlying operating performance.

On the next slide, the quarter-over-quarter view tells a similar story. Quarter-over-quarter earnings are affected by depreciation timing and fair value adjustments related to digital assets. Our adjusted EBITDA in this third quarter of fiscal 2026 was a profit of $5.7 million versus adjusted EBITDA profit of $31.5 million in the previous 2026 Q4 quarter.

In the third quarter of fiscal 2026, we experienced net loss of $91.3 million compared to a net loss of $15.8 million in the previous 2026 Q2 quarter. Operational KPIs, including uptime, efficiency and production remained strong throughout the period. Our internal focus is squarely on cash ROIC rather than accounting volatility. Accounting noise should not be confused with operating performance.

Q3 fiscal 2026 was a solid quarter for HIVE. We delivered strong revenue, expanded margins and maintained a robust balance sheet. Our discipline, fleet expansion and cost control measures continue to position us well to compete in a challenging environment and capture opportunities for growth, both on the hashrate side and on the high-performance computing side in our data centers.

I want to thank our local -- loyal stakeholders and encourage them to continue to follow our dual engine expansion efforts, both in hashrate services and HPC operations.

Thank you, Darcy. That concludes the presentation for today. We will now begin the question-and-answer portion of our call. Analysts on the line, if you could please click raise hand when you are ready with your questions. We will begin to choose and ask you to unmute. Our first question comes from the line of Darren from ROTH. Darren, if you kindly unmute, the floor is yours.

Can you hear me?

We can hear you.

Congrats on all the progress. Two questions, if I may. Just as you kind of push forward on your HPC strategy, can you kind of maybe benchmark how you're thinking about the thought process of returns with AI cloud versus colocation? And maybe what specific metrics, whether it's payback period, return on invested capital, et cetera, that you're kind of making those decisions off of?

And then second question, you mentioned in the, I think, release about New Brunswick, and you kind of mentioned specifically Tier 3 hyperscaler. Is that put in there to sort of benchmark the level you want to build to? Or do you actually have interest from hyperscalers? And I'd be kind of curious about the level of interest there.

Yes. Thanks for those questions, Darren. This is Aydin here. The ROIs typically on the GPUs are approximately 2.5 years after direct operating costs, and we have a lot of experience operating GPUs going back to Ethereum mining days, moreover, having had AI cloud revenue on our income statement for the past 3 years. We had 38,000 NVIDIA A series GPUs, A40s, A6000s, A5000s, A4000s. So we're still running 4,000 of those cards and 34,000 of those cards, we were able to sell at 80% to 90% of face value, and that's what -- those proceeds went to upgrading and buying H100s and H200s.

The point is we don't just talk about it, we've done it. And so we've seen the demand ebb and flow in GPUs, but they have strong residual market value. And so where you're able to ROI in, call it, 2.5 years, but have these cards potentially be worth 60%, 70%, 80% of their value after 3 or 4 years, we've seen a huge uptick in demand for H100s as you've likely heard.

And so the demand comes in 2 ways. One is the hourly rate that the GPUs rent for goes up. But in turn, the market price for people purchasing the GPUs goes up because people realize you can get more cash flow from them. So it's an attractive business, I believe, because if you have the proficiency to do so, if you have the cloud technology platform, which we have and we've demonstrated, and there'll be a lot more updates and exciting news to come as we bring more GPUs online and march towards that 11,000 GPU cloud target, and that crest over that $200 million ARR target in the slides, we believe that it's an accretive business because the residual value -- a, you aligned the GPUs plus you have GPUs that have strong residual value. So you come out well ahead. So I think that answers the first half of your question.

And then the second half of your question, we actually talked about the conversion of New Brunswick in the previous quarter. We bought 32 acres of land adjacent to the site. Engineering design has been advancing since then. And so we have been in talks with groups that are interested. And so there are different ways to deliver power, power shell build-to-suit. And so I can't get into any more specifics other than what we've already disclosed, but a sort of market rate of what gets us about $130 a kilowatt a month for New Brunswick is a secondary -- primary secondary markets. New Brunswick is a secondary market. And so that's where that run rate of approximately $80 million ARR comes from.

You do about 53 megawatts of IT load, but do stand by for updates as we advance our designs and our conversations. We just wanted to acknowledge to the Street that, that is moving forward. And it is not to be forgotten. It's still part of the road map and part of the game plan, but stay tuned for more updates on that. Does that cover it all for you, Darren?

It does.

Next we'll go to the line of Fedor from B. Riley.

I wanted to just like ask about current breakeven price for Bitcoin mining operations, assuming all-in cost to mine, not only power. And additionally, I'd like to understand how Bitcoin at current levels influences your capital allocation decisions for AI and HPC infrastructure. And if you could outline your expected CapEx spending over the next 1 or 2 quarters with any detail on the split between mining and AI HPC investments that also would be super helpful.

Yes, definitely, Fedor. So I think it's quite evident that 2025 is a year of scaling our Bitcoin mining business, having brought on the 300 megawatts in Paraguay, scaling to 25 exahash. So that reflects a lot of capital deployment in that business unit.

And what you'll note from our investor presentation that we just debuted and of course, last quarter, this year, 2026, our focus is on scaling the HPC revenue from $20 million ARR to $225 million ARR, so over 10x. And how do we accomplish that? Expanding the cloud from 5,000 GPUs to 11,000 GPUs, which in my section was detailed growing that revenue from $20 million to $140 million ARR. And then, of course, bringing on the conversion of New Brunswick to Tier 3 HPC for hyperscaler colocation, which at $130 a kilowatt on 53 megawatts of IT load gets you to about $80 million of ARR. So directionally, you can see where the capital deployment is being scaled.

I do want to take a moment to acknowledge though that with OEM vendor financing on our GPUs, we're able to get lease to own, so effectively equal lease payments over 36 months with a $1 buyout, so effectively a finance with single-digit interest rates, which is very, very attractive. Nothing funky like some of our peers have done with pref shares and warrants and all this convoluted mezz financing, just very attractive.

And so we've been able to scale that GPU cloud business, of course, in the partnership with Bell AI Fabric Canada, that data center capacity, we're building the cloud on co-located premises with Bell. So again, that allows us to operate a CapEx-light high-margin GPU cloud business. And so we do have the 100 megawatts in Paraguay that we announced, and we announced that PPA late last year. And so long lead has been ordered, the substation, the design. So that's a long-tail project because, of course, Bitcoin mine economics right now, we're looking at $30 to $35 hash price. So we, of course, are proceeding very carefully.

But what I do want to point out is just remind everybody that we had our recent press release where we're sending nodes to the largest telco player in Paraguay to do a proof of concept for HPC AI. And we're going to be launching GPUs out of an existing Tier 3 telecom center in Asuncion, which is the capital of Paraguay. So we're taking meaningful strides to actually realize and bring HPC compute to Latin America by partnering with an existing data center -- sorry, a telco provider with Tier 3 data centers, much like we found success doing that in Canada with Bell. We're doing it with largest telco player in Paraguay. They're actually owned by a multinational NASDAQ-listed company.

So that is directionally where we're also taking things in Latin America. So that 100 megawatts that we're bringing online, we are really looking at the ability to build a Tier 1 infrastructure today. So the high voltage switchgear, all the power distribution and that infrastructure can be used for Tier 1, i.e., Bitcoin mining or can be expanded upon with chillers and gensets and everything else that you need for Tier 3 for future HPC conversion.

So we're looking at evaluating a road map where we could do both in Latin America. But for right now, we're building the power infrastructure to power that additional 100 megawatts of land. But that is not massively CapEx intensive to buying -- compared to buying ASICs or certainly not building Tier 3. So I would say the biggest CapEx will be building out New Brunswick in -- for converting it to Tier 3. I hope that answered your question.

It does. And just if you allow me to squeeze one related follow-up on Brunswick HPC facility. Specifically, I would like to understand, you already outlined the total CapEx for this project roughly in previous broadcast. But if you can outline current construction status and milestones completed to date for HPC-related portion of the facility? And what's remaining milestones and maybe CapEx for this or next quarter? Just to understand the CapEx spending.

Not sure if we lost Aydin. Fedor, we will follow up you after this.

Yes, no, I'm here. So sorry, the question was -- I don't know, I was put on you for some reason. The question was what are the milestones for the New Brunswick Tier 3 conversion?

Yes, yes. I just can quickly repeat, pardon me. It's just like for -- specifically for HPC portion of this data center, what's already completed to date? And what's the near-term plan with associated CapEx for next 2 quarters, let's say, this way.

Yes. So what we've put out is we've worked -- we've bought the additional land. We're going through the design development. So we have design and permitting underway for that site. The next step would be ordering long lead items. But beyond that, I don't want to provide any more specificity at this time. We'll provide the market with announcements as those milestones are realized. So that's what I got to you right now, Fedor. Good question. I know you want to know more, but you got to hang tight, buddy.

We've got time for 2 final questions. Mike from Northland. I know you've had your hand raised for quite some time. If you kindly mute. The floor is yours.

First question is just for Aydin. If your OEM financing is for 3 years, can you talk a little bit about why you're signing 2-year deals that mismatch? And then secondly, for Darcy, could you help us think about depreciation expense in the next couple of quarters?

Because when you have a longer term, your payments are less. So you cash flow better, Mike. And so we know that these GPUs have great residual value in the market. So at the end of the 2-year term, we may elect to sell them for a gain. We could simply re-rent them out. There's lots of optionality. That's all. But it's mostly -- you just want to structure payment, so you cash flow nicely.

Got it. And then on the depreciation, maybe.

Yes. On the depreciation side, I think you can take a look at what we've got in for the Q3 right now for the 9 months. As we've noted, there was some catch-up depreciation in there -- so if you sort of take the incremental amount that you've got from sort of Q1 to Q3, you can probably take that as running forward. Through Q3, we had all of our ASIC equipment up and running within Paraguay. So that's the best driver moving forward.

Excellent. Thanks, Mike. Gareth Gacetta from Cantor close us out with your final question.

I just wanted to dig in on any potential CapEx requirements for the remaining GPUs you have at Manitoba. So I know you said about 500 have been or will be deployed in 1Q. So I wanted to figure out kind of have the remaining 1,500 GPUs been purchased? And if not, how are you thinking about the funding for those? And lastly, is there any CapEx on the data center side of things at Manitoba?

So the 504, the purchases those were announced in November and the leasing or contract -- contracted term of those GPUs was really announced last week. And so those GPUs should be -- they're being delivered to the facility now and they're expected to go live. So really, the focus is to let the Street know that, that first cluster, 63-node cluster is being commissioned with InfiniBand and all the bells and whistles is going to be live with the client.

This quarter ends March 31. So very soon. We've got 6 weeks left in this quarter. And so I would say, stand by for updates on that. And so once that first cluster is deployed, Gareth, then we intend to reload very quickly. And this model is shaping up to prove to be very successful. And so our intent is to reload and repeat as we rent a cluster, get -- order another one, finance it in a similar way, and get it delivered, rent it out, rinse and repeat. I hope that's helpful. What was the second half of your question?

Just if there's any potential CapEx needed on the data center side of things at Manitoba.

No, no. I mean there was some deposits upfront, but that was all taken care of.

Excellent. Thank you. That concludes our Q&A session and our Q3 2026 earnings call. Thank you for joining. We look forward to sharing more exciting announcements very soon and speaking to you again soon. Thank you.

Hello, and welcome to today's webcast on HIVE Digital Technologies financial results for the 6 months ended September 30, 2025. My name is Nathan Fast, Director of Marketing and Branding at HIVE, and I'm pleased to be your moderator for today's call.

Before we get started on Slide 2, I would like to briefly note the disclosures for today's presentation. Except for statements of historical fact, this presentation contains forward-looking statements within the meaning of the U.S. Private Securities Litigation Reform Act of 1995. Words such as expects, believes and similar expressions identify these statements. Actual results could differ materially, and we disclaim any obligation to update them, except as required by law. For a full discussion of risk factors, please refer to our most recent SEC filings at sec.gov.

In addition to discussing results that are calculated in accordance with GAAP, we will also reference certain non-GAAP financial measures, including adjusted EBITDA, adjusted net income and free cash flow. Management uses these metrics to evaluate operating performance and believes they provide investors with additional insight, and they are presented for supplemental purposes only and should not be considered in isolation from GAAP results. Reconciliations to the nearest GAAP measures are included in the appendix to this presentation and in the press release and Form 8-K furnished to the SEC.

On the next slide, I'm pleased to introduce today's presenters, Frank Holmes, Executive Chairman; Aydin Kilic, President and CEO; and Darcy Daubaras, Chief Financial Officer. I would now like to hand the presentation over to Mr. Frank Holmes from a macro recap of the quarter. Frank?

Thank you, Nathan. Let's get into my macro recap, and I'm going to try to weave in some of the most recent past couple of weeks' sentiment, unpacking the drivers of short-term negative sentiment in data centers and AI infrastructure, which I believe is just setting up for a phenomenal buy for us an industry whole. I'm going to try to explain this in this presentation and why HIVE is so uniquely positioned in this macro theme. Next, please.

Well, HIVE's journey to really accelerate its growth in Bitcoin mining. It was stalled after Putin invaded Ukraine and sources of hydro and geothermal electricity around the world all of a sudden went shut down or very small. There was no scaling ability until the opportunity a little over 18 months ago when I reached out to the President of Paraguay and flew down to meet him because they have surplus electricity. I'm going to try to walk you through sort of this journey. And that this President is really quite unique. President Santiago Peña is a former rugby player. He is 47 years old, studied at the University of Columbia, New York, worked for in Washington with the IMF and has had an incredible journey in Paraguay and building the country and becoming recognized. And every time I go down, I find something new that I love more about the country.

But the country is very unique and very pro-America. So to give you an idea of all those negative news about Latin America, with Valenzuela and some of the drama of the Panama Canal and then we have, for sure, Columbia's President. I think it's really important to look at how Communism and Socialism has penetrated deeply into Central and South America, but not in Paraguay. And I think that that's what's really important for investors and why we felt so confident this is where HIVE would scale its Bitcoin operations, sourcing green energy, surplus energy and really also helping out the overall economy grow, and I'll show you some of the other unique things we've done in schools that are nearby us.

But President Santiago Peña was the sole Latin American Head of State present at the 2025 Gaza Peace Summit. So this is something else that's really important who's pro-Taiwan. And one of the largest suppliers of food commodities for Taiwan, I believe, for a population with 6 million people, they produce about -- food for and which they export for 60 million people. Next, please.

So here's our President of the country, Gabriel Lamas, who's an electrical engineer, worked for the government's utility company. As you can see, he's very tall compared to the President who's like 6'2, Gabriel is like 6'4, 6'5 and he's just been a great leader of maximizing the whole scaling, the concept of scaling and coordinating and working with our CEO, Aydin Kilic, who'll speak in a few minutes. And Luke Rossy, who's our Chief Operating Officer; and Bill Gray, our CTO. Gabriel has done a phenomenal job accelerating the scaling from 6 to 25 -- the journey to 25, we've announced for 24, and we're very close here to 25. So we feel very confident of hitting that.

So you can see here, Gabriel was recently with them. And there's a picture when I first met the President a little over 18 months ago. And the economy is one of the strongest and steady growing economies in Latin America. So that's another real positive sign for the country. Next, please.

As a money manager, I have always say that government policies are a precursor to change. And Peña's policies for fiscal development are outstanding and he has pushed to make sure that every child is fed and then upgrading the school system, but we're helping with that. And we're helping, especially in rural areas. And so this story, what I'm trying to point to is that Paraguay is powering high velocity ascent to 25 exahash.

And what's interesting is that yesterday -- sorry, 2 days ago, is November 12, and that is Hayes Day because in 1870, there was a tragic war, a trilateral war with Uruguay, Argentina and Brazil trying to take over this massive green belt country, and they wiped out genocide all men and boys to the age of 10. It was tragic, and it was President Hayes that came in that resolved that they kept about 60% of their land. And to this day, the country celebrates November 12, Hayes Day and the largest state in the country I'm pointing to here is President Hayes. So there's a very strong bond with America that when I talk to investors, the average America is not aware of. Next, please.

So part of that journey in addition to rapidly scaling and building out data centers, campus across the country, building a Buzz is also -- it's creating lots of local jobs in construction. We've shown you a visual of lighting the streets in Valenzuela, a city near us. But then it's also for the kids, and this is a thank you from a former leader, a mayor of Valenzuela for lighting up the streets and making the street safe at night for women and children. Next, please.

So when we went out and saw the schools, which are very close to us, I'm talking about 0.25 mile, 0.5 mile away. We started in this process of upgrading these schools and putting in air conditioning, new electrical painting, new chalkboards, infrastructure and you can see here a PlayStation here for kids so that they can go and play and enjoy these facilities and also for embracing data centers. There's a big, great disconnect even in very advanced countries like Sweden that take it for granted all the apps on your mobile phone and Netflix and Spotify. They think it's all free, and they don't realize no data centers, no digital realm that we all live in and we rely on. So this is sort of a visual of relating this to the children of Paraguay. Next, please.

You can see the picture, this is the key people here. You can see Gabriel Lamas is in the middle, and he's kneeling down because he's so tall. But behind is myself, and you can see to my right, Luke Rossy, and you can see to the left, is Aydin Kilic, our CEO, who's also electrical engineer like Gabriel Lamas is. And beside Gabriel is Craig Tavares. Craig is our President of Buzz, which is building out our AI business. And you can also see Johanna Thörnblad, who's the President for Sweden. And you can see one of our independent directors and other key people that have been very much involved in the growth and acceleration. Investment banker, Jamie Brown is there. So it is a lots of key people that are helping drive this.

But behind the visual is the largest, biggest hydro dam in the Western Hemisphere. It's 5 miles long. It's epic of size, producing over 10 exahash of electricity, 5 for Brazil, 5 for Paraguay. Paraguay is a small population. So they've been selling some of their electricity to Argentina, which has been only really horrific for their fiscal management of a country. So it's a small country, and that's what's good about Bitcoin mining. Not only do we build this infrastructure, you need substations from this high-voltage electricity, and we build those. And then we build a campus of data centers, and then we end up hiring 50 young engineers that there's no real job opportunity for them. And -- but now there is with HIVE. And so the -- it's really, for us, an exciting part of our journey of creating this from little kids' education to creating career paths for young engineers. Next, please.

So on that journey of going to Latin America, I really want to show you that this year has been a phenomenal year for Latin America of 40 ETFs, so the 40 biggest market cap stocks versus the S&P. They've crushed it. They've really outperformed the overall S&P 500. And I think that, that's what's really important for investors to recognize. Next, please.

So on this big build-out, I call it an arms race. China has significantly surpassed America in building data centers. But before that, they build power. And the power came from hydroelectricity. And they got the water from the glaciers in the Himalayas and rerouting water from Tibet down to creating dams after dams and building out a huge grid. And America has to get caught up, and it is. And under the stewardship and leadership of President Trump, as soon as he came to power as the CEO of NVIDIA says that impression the most about Trump is that he embraced the need for AI and power and data centers. And here is a classic where Bitcoin miners were the first to go. This is in West Texas, Sweet Abilene, Texas, where Crusoe was using flare gas, surplus energy from solar, wind and flare gas today is the largest CapEx spend in America. And I'm told in the world right now in fast tracking a $500 billion AI data center vision from former flare gas to a Bitcoin miner.

I'm really trying to highlight that, that Bitcoin miners were the early pioneers of building the necessary infrastructure wherever there was surplus or additional or wasted electricity and creating taxes for the local economy, creating jobs for the local economy. This particular site here created 2,500 construction jobs. And when that's all done, it will be 500 jobs and a big boom to the local community. Next, please.

So Bitcoin mining. The goal is up 2.8% of the global hash rate. We really believe the blue-sky opportunity in Paraguay is probably a gigawatt of electricity. That's the potential as -- when you use the word blue sky and not to go into the moon. That's way out there when you talk about blue sky, but Paraguay, no doubt, it just makes the perfect place for us to expand our footprint. Backed by the most consistent and stable executive leadership team in the sector, I'm very proud about HIVE. HIVE's CEO is the longest standing. CFO is clearly the longest standing. The turmoil in the industry has been just phenomenal. And we've been able to do this and have operating income no matter how volatile Bitcoin has been. We have always been able to squeeze out operating income and have huge expansion, which we've done this year in -- with this growth, the most efficient Tier 1 Bitcoin mine data center operator, that's really important for you to recognize spanning 9 timezones in 3 countries and 5 languages. We have the lowest industry SG&A per Bitcoin mine and best-in-class for uptime and efficiency.

So we know a lot about Bitcoin mining, and we also know in depth and breadth more than these other bitcoin miners that are going into the space of AI. We've been doing it for 3 years. I believe we're quite successful. It's much more complicated. It's much more detailed and substantially more expensive per megawatt per chip to go build out a Tier 3 HPC or high-performance computing data center for AI. Cambridge University has said, it's 33:1 from a Bitcoin mining. So what Bitcoin mining does is it sources electricity, builds up the necessary infrastructure and then you go as further things are developed in an economy such as dark fiber optics to move the data along. It takes time for that in many emerging countries. But for Paraguay, I have to agree with Secretary of State, Rubio, of saying that Paraguay has the potential to become the data center capital of Latin America. And we hope to be a significant part of that HPC build-out as we are doing what we're doing now in the Bitcoin mining. So we have this proven track record. We're now scaling in AI, and let me walk you through a few comments on that. Next, please.

So November 10, HIVE reports October production, 289 BTC, achieves the 24 exahash and completes Paraguay expansion while fueling Canada's Tier 3 and AI data center growth. Next, please.

We've shown this growth by buying a former Bitcoin mining data center within the city of Toronto near the airport. So uptime is all the time. And going from Tier 1 to Tier 3 will be much faster than starting from scratch. This is a perfect ideal location, and we feel that this will be a 2-tier payback. And clients, we've already built out 10,000 customers around the world, but the unique relationship we have with Bell Canada is transformative besides going from Tier 1 Bitcoin data center to Tier 3, which is much more expensive because you need much more air conditioning because you actually build [indiscernible]. When you think of these H200s and Blackwells, these chips are more intensive consumption of electricity. So you need much more air conditioning and the HVAC, and you need more support. So those dynamics are really important for investors to realize besides headlines, it takes time to do it. We've done it, and we're now expanding rapidly. Next, please.

So when we look at our suite of assets around the world, one of them is that we bought an asset besides Toronto. We bought it in Northern Sweden, Boden near Facebook, and will convert a data center there like we're doing in Toronto. And that's -- that will go through -- it's already started that process, that conversion. But then we said we have New Brunswick on the border of Maine, and this is a beautiful campus of data centers. So we made a press release just recently, we bought more land, secured the position to build out our Tier 3, and it will be the biggest pure HPC data center complex in Canada at this time.

So we're very excited about this. It's a tremendous blue-sky opportunity. As we're going right now for this particular asset, first is to tie up the land. We have the electricity and now to get the engineering, complete drawings so we can start the construction in the first quarter of next year. Next, please.

So Buzz high performance overview, HPC. So data center locations right now today is downtown Montreal and Stockholm, Sweden. We have GPU clusters. We have over 5,000 GPUs and AI cloud services built -- purposes-built AI cloud with managed services in agentic refinery. That means that you can provide more services, especially when you have 600 sales force at Bell Canada that needs to get caught up quickly, and we're the strategic partner for that. We've already shown what we can do. We're doing $100 million in revenue, and now we're going to scale it over the next 12 months, fivefold.

Blue sky, well, just showed it to you, we're going to go through a transformation of Tier 1 to Tier 3 HPC data centers. And I think when we look around the world right now, the back of the envelope is 90 megawatts. And if you start looking at other assets we're working on securing, it's much higher. And the long-term vision and opportunities for Paraguay are just immense. So we're very excited about this growth opportunity. But unlike other people, we've done it, and we know what it is to build data centers. We know what it is to scale Tier 1 data centers, and we know what it takes to build Tier 3, and we are scaling. Next, please.

This was a very important transaction for Buzz HPC in Canada because Bell Canada is the largest telecom in the country. This partnership is to help build sovereign AI infrastructure through the Bell AI Fabric initiative, advanced NVIDIA hardware and networking, deploying NVIDIA, Ampere, Hopper and Blackwell GPUs with Quantum-2 Infaband (sic) [ InfiniBand ] across Bell's fiber network and data centers. Canada is a huge opportunity because they're way behind, highly educated, and they've just been slow but under Prime Minister Carney. This has now been elevated to the very top with the AI minister. This is full tilt on the country doing everything to get caught up, just like America is getting caught up to surpass China, tremendous initiatives.

So this is very positive and constructive. And the initial deployment and national reach has been 5-megawatt GPU facility in Manitoba with plans for the national expansion, like we talked about Eastern Canada. And then data sovereignty, security and sustainability has become a very important issue for many of these countries. So we believe that HIVE has his head office in Texas reporting for Nasdaq and doing financial reporting under GAAP. But the intellectual capital and the Buzz are their Canadian companies that would qualify under the sovereignty of Canada and wouldn't be a challenge to the degree that people are worried about like in France over the Patriot Act. So I really think some of this becomes really important to recognize how uniquely HIVE is positioned for this. Next, please.

There's our growth. February was $13 million. May was $20 million. And we're hoping that to get these others deployed as fast as possible to get us up to $40 million in the fourth quarter, which is going to have this run rate, which is going to be, for us, our year-end is March.

And then going into '26 based on all the projections we have from hooking up our suite of NVIDIA chips, looking at those contracts, it looks like a very conservative run rate of $100 million annual run rate with extremely attractive margins for growth. Next, please.

So HIVE's top institutional investors are Citadel Advisors, Schwab Crypto Thematic ETF, Amplify Investments, Tidal Investments and Charles Schwab as a discount broker. I remember when -- a few years back, it was used to be Robinhood. And -- but now it's showing up with Charles Schwab, and Charles Schwab bought TD Waterhouse. So it's just great because they are the biggest beast now. And it's great to see broad retail in addition to institutions that are here and met with the fund managers of most of these companies this week in a conference that Aydin Kilic, our CEO and I attended in Miami that was put on by Cantor Fitzgerald that was called crypto, infrastructure, AI and energy, big, big theme going forward. Next, please.

So we celebrated 4 years on Nasdaq, and it was just a major home run. We had 100 people out for it. We just -- the ministers, 2 ministers from Paraguay and New York City came out for it. I see the President of Bermuda is there, Tim and I see auditors. I see just a huge crowd, and we want to thank everyone for helping us celebrate that. And there's Adam -- sorry, there is Aydin hitting the bell and that's my daughter. And you can see Craig Tavares, the President of Buzz, his daughter. So it was a home run for us all to be participating.

But what's happened short term before I pass you on to other smarter executives is sort of AI adoption. There's been some really important research has come out from Morgan Stanley that's been circulated about framing the AI boom. And I'm a big believer that the AI boom is for real. It's there. But sentiment drives a lot of these markets short term. And all you have to do is get these clusters of calling bubble. There are so many analysts who got a PhD, they say in bubbleology, but they do not really understand what is driving the secular market. And we're due for this correction, which we're getting, which only sets it up for a more attractive buy. If we were valued at other data centers, seeing these other companies in the Bitcoin mining that do not have any high-performance computing facilities today, do not have any AI revenue generating for their NVIDIA chips have valuations that if you apply to HIVE, HIVE would be a $20 just to give you an idea of sort of the great opportunity that we have as we continue to scale.

And I think that it's important to just recognize what a lot of this noise that's taking place is short term, and it gives you just this opportunity that you see CoreWeave go through a correction. It's just part of the delays and disappointments on the infrastructure build-out that some of these people are experiencing, but it's not going away. This is an arms race, just like the push by NATO in Europe and Canada and the U.S. of huge military spend, huge rebuilding of all the military and that's just not tanks, but it's tanks with AI, with GPU chips, and they need data centers. And the digital economy is going to continue to need and it's going to continue to grow.

We have these things they call phantom demand. We experienced with the Bitcoin mining when the Chinese shut down Bitcoin mining and everyone from China started making these incredible bids for energy all over the world. They were just phantom, trying to see if they could secure some energy and it comes up. But it's -- I just think that my history of these cycles is that we are still very early innings on the adoption. And when you look at a pure data center and the multiples they trade at to revenue and the multiples they trade at to cash flow really makes some of the Bitcoin miners outstandingly attractively priced. And in particular, my biased position is with HIVE when I look at multiples to revenue as we are powering forward with our twin engine turbocharged strategy of both Bitcoin mining growth, not only this year, the biggest growth, we also plan for next year. When I look at our peers, we have the biggest growth profile for Bitcoin next year. And we're still very bullish about it.

But at the same time, we're pretty conservative of how we write down the value or depreciate our ASIC chips faster over 2 years because every 4 years, the supply halves. And what you're seeing, which I remain very positive like Moore's Law is that you are witnessing the technology of energy efficiency that going back 10 years ago was like 1,000 joules per second consumption of energy for a chip -- ASIC chip. And now it fell a couple of years back to 30. And now they're coming in at 12 and people are working on 8 and 6, so by the time of this next halving, I think there'll be a 2 and 3. We're just with AI going to be able to build more energy-efficient data centers, more energy-efficient ASIC chips. And that's only going to mean less consumption for the global Bitcoin network that's decentralized, but it means that you're still going to be a very profitable business and being significant in laying the railway tracks for the growth in high-performance computing. Thank you.

And thank you, Frank, for that executive overview. Frank, of course, is the Co-Founder, Executive Chairman of HIVE. I'm going to get into a bit more of a deep dive on our accomplishments for this quarter and our strategy for 2026. Let's hop into it.

So it was a record-breaking quarter for HIVE. $87 million of revenue. Of that, $82 million from Bitcoin mining and over $5 million from our HPC business. So we're a $20 million ARR for our AI cloud business, which is phenomenal and a $42 million gross operating margin for the quarter. That translates to a $31.5 million adjusted EBITDA and a $23.8 million EBITDA with $38 million of depreciation this quarter that leads to a net loss of $15.8 million, which, of course, is on a noncash basis. But moreover, what I'm really proud of is that we realized an 18% ROIC this quarter on an annualized basis. And finally, as of September 30, we had 210 Bitcoin in the treasury, although in addition to that, we have 1,992 Bitcoin pledged. So between our pledged and what we have in the treasury, that's about 2,200 Bitcoins. So it's been a phenomenal quarter, and I'm really proud of how our balance sheet deployment to scale the business was, in my opinion, a great success. Let's go to the next slide.

We are very serious and very dedicated about having best-in-class ROIC as the numbers this quarter once again affirmed, and it comes from disciplined capital allocation. We always focus to get the best ROI, we buy ASICs. ASIC purchases are the biggest variable in realizing an ROI and then from their free cash flow investments in the Bitcoin mining business. We're experts in that. And I believe pound for pound, we're the best in the industry at that. We also lead with best-in-class uptime. We run our ASICs to their entire economic lifestyle to maximize free cash flow and profit on our Bitcoin mining business from the investments we make. And by the way, we also lead the sector with corporate G&A, and we have the numbers to back it up. Next slide.

The overview of the global business. Currently, we're approximately 10 Bitcoin a day with 25 exahash capacity installed. And this was fueled by a transformative growth in Paraguay, where we grew our business from $100 million from fiscal 2025 to currently, we're at $400 million ARR. Additionally, we have 5,000 GPUs between Canada and Sweden operating AI cloud, and that is currently at $20 million ARR as per our quarterly financials. But the really exciting news is our target to grow to over $200 million ARR between both AI cloud and also hyperscaler colo. And finally, we talked about our capital deployment strategy, prioritizing ROICs.

So a double-click on our BTC pledge. So we have great downside protection as well as upside. So the -- almost 2,000 Bitcoin that we pledged, 1,234 Bitcoin were pledged at 87,000 -- and so what that does, it gives us upside with Bitcoin today at approximately 297,000. Even with Bitcoin at 97,000, there's over $12 million of upside in our currently pledged Bitcoin. And moreover, we additionally had pledged 758 Bitcoin at an average of about 115,000. And so that's $14 million of downside protection from the other component of our pledge.

So collectively, if you look at the downside protection of $14 million and the upside potential of the $12 million, even with Bitcoin having corrected where it is, the value of our pledge has been accretive to the tune of about $25 million, which I think is very phenomenal. And it just shows that we're very methodical and thoughtful. We've use our cash flow from operations. We used our HODL and earlier this year, some of the ATM to scale the business to get to that $400 million ARR, and then we have some very exciting things in the pipeline for the AI and HPC business with Buzz. Let's hop to the next slide.

So the dual engine strategy of growth. Cash flow from the Bitcoin mining business helps us to scale Buzz HPC, where we have 3 AI data center expansions going to Tier 3 plus, and we'll get into that. But right now, at glance, the Bitcoin mining business, 25 exahash, a blended fleet efficiency of 17.5 joules a Terahash. We're at $400 million ARR, as we mentioned, approximately 50% operating margin to electrical costs based on current mining economics and it's a footprint of 440 megawatts of green energy.

Now the exciting news we recently announced, in addition to this, we have an additional 100 megawatts that ANDE and the government of Paraguay has approved for HIVE, which we're very exciting about. That gives us the potential to get to a capacity of 35 exahash for the end of next year, and that would bring our fleet efficiency down to 15 joules a terahash if we were to upgrade with the latest generation machines now.

We only buy ASICs when it satisfies our ROI targets. So we're focused on building the infrastructure. And the other really great piece of news is that this is at the Yguazú site, which we purchased from Bitfarms and originally was designed as a 300-megawatt site. So what we're really excited to announce is that the incremental cost to bring that 100 megawatts on is only $250,000 a megawatt. Recall that the all-in purchase price for the site and even to build our Valenzuela site is $400,000 a megawatt all in. That's land substation buildings.

Well, the good news is that all the civil work is already done for the additional 100 megawatts at Yguazú. So all we do is we drop in the transformers for the substation and then the hydro mining infrastructure, $250,000 a megawatt, very incremental and attractive cost to bring online the capacity. And with that capacity, we have optionality. We can either install next-generation ASICs. Of course, we always shop for ASICs at about a 1-year ROI or we can look at other strategies to monetize that 100 megawatts of green energy. And we have ordered the long lead time items, namely the substation, which are expected to arrive in Q1 of 2026. So standby for updates on that. And by the way, if we were going to do 35 exahash, the current mining economics, that will be about a $550 million ARR for end of Q4 2026. Moreover, let's talk about Buzz.

We have a very exciting year ahead for Buzz. We've recently announced that the additional 6,000 Blackwell GPUs, B200s for H1 and B300s for H2 will come online in 3 data centers, bringing our current 5000 GPU cloud to a total of 11,000. We're going to double the size of our GPU AI cloud by the end of 2026. And moreover, with the conversion of New Brunswick, we bought additional land. So that campus is now over 32 acres -- sorry, over 30 acres. And it's currently an operating tier 1 70-megawatt data center. We'll convert that to Tier 3, and that will provide us some hyperscaler colo. You add that all up. It gets us to a target of $225 million for the end of 2026 for Buzz. And collectively, our target for the end of 2026 can be $750 million for both the dual engine strategies, of course, we prioritize ROIC. And so on the scaling of the Bitcoin mining business, we will scale that if the ASIC investments allow for ROIC and they're attractive. And of course, on Buzz HPC, we will be bringing online these GPUs in our partnership with Dell for the first phase of our Bell partnership. Let's go to the next slide.

So at a glance, the global footprint for Buzz now, this is just HPC. We currently have approximately 5,000 Hopper, which means a H100s and H200 series GPUs, about 4,200 A-Series GPUs. And those are in Sweden and Quebec. And so the additional 6,000 Blackwell GPUs that we're going to bring online are coming on in 3 tranches. One is through the partnership with Bell in Manitoba. One is through the Toronto data center, which we recently closed on the purchase of and announced the conversion of Tier 3 liquid cooled. And then, of course, our facility in Boden, Sweden, which will also be Tier 3 liquid cooled.

So Toronto and Sweden are owned -- HIVE and Buzz owned and operated data centers, we own the land and buildings, and we'll be converting those to liquid cooled. Those will come online in Q4 of next year and each facilitate 2,000 NVIDIA Blackwell Series GPUs. And so this shows you the math and the footprint of how we scale from 5,000 GPUs to 11,000 GPUs in our cloud business.

And then, of course, with New Brunswick, the conversion of that to a Tier 3 colo with expected IT to deliver possibly 50 to 55 megawatts market rates in that sector would be about $85 million of additional ARR. And so this gives you a high-level glance of how we incrementally scale our $20 million ARR to date to over $200 million target for the end of next year. Next slide, please.

So beautiful photo of one of our H200 clusters operating. Now to help clarify, on the GPU business because we're doing both GPU, AI cloud as well as -- to target hyperscaler colo in New Brunswick. So if we focus on the GPU cloud business, the growth is $20 million going to $140 million, and that reflects the growth from 5,000 and to 11,000 GPUs by bringing on an additional 6,000 Blackwell GPUs in those 3 facilities.

And currently, we have a good mix of long-term contracts for 3 to 12 months, where we're monetizing our $20 million ARR as well as on-demand, we have over 10,000 unique users through our partner networks. And I think the really exciting thing is to highlight that we're focusing on liquid cooled Tier 3 data centers for the future as these NVIDIA reps get more and more dense. We've heard that the Blackwell GPU clusters for B300 could be as dense as 180 kilowatts per rack, and rumors are that ribbon could be as much as 200-plus kilowatts per rack. So it's building for the future. And of course, we're really excited about that partnership with Bell. Let's hop to the next slide.

So this was released this quarter, a very exciting initiative where Bell is Canada's largest telecom provider. They have partnered with Buzz HPC, and the team has done a phenomenal job to bring this deal to fruition. And we've just executed all the operating agreements. And so we expect that the first cluster of B200s, which you'll be operating in Dell servers to arrive at the end of this year and go live in January of 2026 and bringing online that initial incremental ARR, and this will be part of the sovereign data strategy, Canadian owned, meeting the residency requirements. Bell is a -- sorry, Buzz is Canadian company with a great pipeline of data centers in Canada through Bell and the Toronto data center. So this is all very exciting. Craig Tavares has done a phenomenal job advancing that initiative, along with Mario Sergi and our HPC team. Next slide.

So by the numbers, again, on a time line basis, this shows the current $20 million ARR. And as we bring those incremental clusters of B200s into the Bell partnership, the 1,000 for Q1, another 1,000 by Q3, each cluster of 1,000 GPUs on long-term rental basis. So we're forecasting this, if we just go to your long-term contracts and you go up market rates on semi analysis to see where H200s and B200s trade-up. So this is based on dollar per GPU hour. Long-term 2-year contracts is about $20 million per cluster of 1,000 GPUs. So we're at $20 million today. If we bring on 1,000 GPUs in Q1, it gets us to $40 million ARR. Another 1,000 GPUs in Q2, Q3 gets us to $60 million ARR. Now the big ramp is Q4 next year because that's when the conversion of Toronto and Boden data centers will be complete. Each of those will bring on 2,000 more Blackwell GPUs. By Q4, we target it will be B300s. And of course, a cluster of 4,000 GPUs would add about $80 million ARR. And so that's how you get to the $140 million ARR growth on the AI cloud business with GPUs.

And then in addition to that, separately would be the hyperscaler colo with a conversion of New Brunswick, currently a 70-megawatt Tier 1 into a Tier 3 plus data center. That incremental revenue would be about $85 million based on $130 kilowatt a month. And so this is the growth strategy for Buzz for 2026 with both AI cloud and hyperscaler colo. We have the land, we have the power, we have the expertise in running the GPU clusters and we've been doing it for the last 2 years. Next slide.

We also are vertically integrating the tech stack. So HIVE and Buzz, we have expertise in building and operating data centers. We built expertise building and operating the GPU clusters. But especially with the partnership with Bell and sovereign initiatives within Canada, where we want to sell managed services to AI enterprises, it's about vertically integrating the stack and having a platform that enterprises can use or sovereign clients. And so this means custom endpoints allowing for agentic AI, inference training, tuning. And so this whole platform, we're very pleased to announce. And the next slide is Buzz HPC cloud.

And so this is just launched, and this has been pressure tested. This is scalable. This allows us to sell sovereign compute for our clients in Canada, and we think that this is a very exciting development that really opens up the sovereign market in Canada, which, of course, Canada is the home to the Vector Institute, the home of Geoffrey Hinton. Our buzz team led by Craig Tavares is doing an amazing job working with the different strategic partners in Canada with Bell, with Dell, with the different research institutes, and we have some exciting news in the pipeline. So stay tuned for that.

But moreover, what this does is with our platform selling AI enterprise solutions, it's a much higher margin business. Let's go to the next slide.

And this is the money shot. The Buzz cloud platform has been ranked and awarded Bronze by SemiAnalysis. This is the benchmark in the industry that looks at all the neo clouds. And you will notice, and I'm very proud to see our team has worked so hard. The Buzz team has done such a phenomenal job on our first go around with SemiAnalysis ClusterMAX, we received bronze.

You will notice the other industry peers. And when I say industry peers, I'm referring to former Bitcoin miners that are now pursuing HPC AI cloud. The other companies at our peer group are all in the underperforming category. So -- and by the way, even below that, you'll see some names that have raised a lot of capital in the unavailable. So they weren't even able to really complete the benchmark test. And then the next category is underperforming where they completed, but obviously not up to the standard. And then beyond that, which is where Buzz is, among some other very reputable clouds is Buzz HPC. And our team believes that we're very close to being able to get a silver qualification next year, which would put us right up there with AWS. So super exciting. And we're very proud of this accomplishment. And again, Craig and the team have done an absolutely phenomenal job on this. Let's go to the next slide.

So with the blue-sky potential integrating the cloud platform to allow us to sell managed services for AI enterprise clients, we're looking at a much higher margin business. And so instead of renting GPUs for bare metal at long-term rates, which is what we presented on the previous slide, and that's the base case. The bull case or the blue-sky case here, if all the GPUs were monetized using our cloud platform for AI enterprise clients in Canada under the Buzz sovereign strategy, it would be a much higher margin. And by the numbers, you would be looking at approximately double the dollar per GPU hour unit rate.

And on the previous slide, we were basing it on about $2.30 per GPU hour and these figures here on $4.50 a GPU hour, you get about $40 million of ARR per cluster of 1,000 GPUs. So what that does is if you bring on that first cluster and they're all being monetized through the Buzz cloud platform, it could bring an additional $40 million. That would mean the current $20 million of ARR goes to $60 million ARR and subsequently increments to almost $0.25 billion ARR just from the AI cloud business.

Again, this is a blue sky, not necessarily all the GPUs will be sold through the cloud platform. For our sovereign clients, we still have existing relationships where we do sell AI compute globally. And so we believe the margin will fall somewhere between these targets and the base case targets.

But if you add on the capacity of hyperscaler colo by converting all of New Brunswick, then that brings you to the north of $300 million ARR target. So if you look at this blue-sky and the base case, I do think that getting into the $0.25 billion ARR target for end of next year is realizable potential, and we're excited to take it to the next level. And we believe that if you look at where the multiples where some of the peers are trading, it easily puts our AI cloud business at a multibillion-dollar valuation if we were doing $250 million ARR. And by the way, this is much higher margin. When you're selling through cloud, the margins are about 90%. And in the previous -- the base case where we would do long-term GPU contracts, at least for Blackwells about 80% margin after data center and electrical costs. So we're very excited to present this and 2026 will be a phenomenal year.

The other exciting piece is that to bring these GPUs online, we have vendor financing. And so we actually are not required to raise capital to bring on this cluster of GPUs because 6,000 Blackwells will be about $360 million. And we're bringing them on. We're doing about 1,000 GPUs at a time. And when we do this, we rely on vendor financing, lease-to-own which allows us to, therefore, be very mindful about our ROIC. And really, the CapEx is just to finish the conversion of our Toronto data center and our Boden data center. And those 2 facilitate the growth of 6,000 GPUs. Moreover, the colo partnership with Bell, it's colo, and we're getting a very attractive rate, doesn't require any significant CapEx.

So to fund the growth to bring online to 6,000 GPs does not require any significant CapEx beyond what I believe we can fund from our current balance sheet. And that's what's really exciting. The conversion of New Brunswick to HPC would be a more capital-intensive project, but stay tuned for updates on that. Next slide.

So on the Bitcoin mining side of the business, very proud to announce we're mining approximately 10 Bitcoin per day, even with difficulty at 152 trillion, currently operating a footprint of 440 megawatts of green energy. We have a proven track record of scaling now as we've demonstrated industry how quickly we've brought on this capacity. Again, we've maintained not only best-in-class uptime and efficiency in the sector, but lowest G&A per Bitcoin mine, which you would expect because we've done this with a very lean team as we've quadrupled our revenue this year.

We've done so by adding a small number of additional key hires. Of course, our efforts in Paraguay are led by Gabriel Lamas and Carlos Torres, who have both done an absolutely phenomenal job. We're extremely proud of our Paraguay team. And I'm going to be flying down there in a couple of days for my quarterly site visit. I always like to visit every data center, walk the fields and walk the hot aisles, walk the cold aisles, just like to feel the energy. And of course, we do this all with disciplined capital allocation, which we prioritize ROIC, and we're going to double-click on that later in this presentation. Next slide.

So Thanksgiving came early. We did target that we would hit the 25 exahash and completed 300 megawatts by U.S. Thanksgiving. And we hit it a couple of days ago, which was really exciting. We put out a press release around November 11. And so Thanksgiving came early. So that's exciting. And it's 10 Bitcoin mined per day approximately and 25 exahash installed ahead of schedule. I think we've really demonstrated to the market that we have expertise. And I think for the longest time when the market was watching high, why aren't they scaling? Because we wanted to scale when it was accretive. And by the way, finding green energy, obviously, is more challenging than if you just look for any type of energy. So very happy that the team delivered. Everyone has done a phenomenal job. Shout out to Luke Rossy, our COO; and Bill Gray, our CTO as well. So the whole team, Darcy, Gabe, everybody has just done a phenomenal job. Next slide, please.

So this is just a scenario analysis that shows you we've seen Bitcoin correct to just under 100,000. But of course, with Bitcoin at different incremental prices, you could see that our current annualized mining margin is about $200 million with Bitcoin 100,000, but that could easily get to $300 million at Bitcoin at 125,000 and at 150,000, almost $400 million of annualized mining margin after electrical cost. This is in not revenue. This is mining margin after electrical cost. So $200 million to $300 million to $400 million is as measured against Bitcoin price, really at 100,000, 125,000 or 150,000. So a lot of upside in the Bitcoin mining business, of course.

And so just a quick note on that. When we look at some of the hyperscaler colo deals that have been done, Cipher announced a deal with AWS $102 a kilowatt. If you actually dollar denominate that per kilowatt hour, it works out to about $0.14 a kilowatt hour. So with Bitcoin at about $43 hash price, Bitcoin was mostly at $102,000 this week. That's about $0.12 a kilowatt hour with our S21+ Hydros. And so you're actually not that far off on a dollar per kilowatt hour basis compared to hyperscaler colo.

Of course, what institutions like is the long-term stability, a fixed price for 12 or 15 years, you can underwrite, put a multiple on, and we're very much aware of that. I think the Morgan Stanley report on HPC did a good job spelling out that strategy. But what you have in the coin mining is upside because when hash price was $55, the mining business was actually doing $0.15 a kilowatt hour, so even more top line than the Cipher -- hyperscaler colo deal. But what's interesting is that, of course, our bottom line, our operating cost is about a $0.05. So you've got a much higher margin business.

And that's why we're able to target 1- to 1.5-year ROI in the Bitcoin mining business, whereas your ROI when you're doing a hyperscaler colo, if you're spending $8 million to $10 million per megawatt to build a Tier 3 data center, you're looking at roughly 8-year ROI. And so -- but we do both. We're targeting the HPC business with the AI cloud. That's about a 2.5-year ROI on the GPUs. And we're very much aware that institutions like the long stable cash flow of hyperscaler colo, and that's why we're very pleased to bring to market, the New Brunswick opportunity and put that in the 2026 pipeline. Let's go to the next slide.

So just a high level for all the research analysts out there, where the hash rate is and how much power is online. And of course, the Phase 3 in Yguazú, the additional 100 megawatts actually gets us to a global capacity of 540 megawatts. Of course, 440 megawatts today is operating hydro power. And might I add, again, I just want to reinforce, we are not committed or we do not have an immediate capital need to bring on 10 exahash of ASICs. We will do so if economics and ROI satisfies our principles of realizing a 1- to 1.5-year ROI after cost for the ASICs. And there are other ways to monetize that extra 100 megawatts of capacity as well, by the way. So stay tuned for that.

We've got a lot of exciting things for 2026. We're putting all the fundamental pillars in place for a year of growth and to manage both upside and downside. And that's how we've lasted for, coming on 8 years now, being the first public crypto miner with the longest standing C-suite might I add between Frank, Darcy and myself. But let's hop to the next slide.

So -- by the numbers for those that would like a snapshot of what 35 exahash would look like on an ARR basis. Well, as we mentioned, currently, we're at $400 million ARR, doing about 50% margin after electrical cost, $200 million of margin after electrical cost on the current Bitcoin mining business for 25 exahash about 17.5 joules per terahash efficiency. We were to scale that to 35 exahash with latest generation ASICs throughout 2026. And by the way, we do believe that ASIC prices for the new generation here will come down and be more attractive. And so we would evaluate those during the spring to summer of 2026 as the 100 megawatts would be completed around August to September 2026. So we're in no rush. But with that 35 exahash, current mining economics, which is modeled on $42 hash price brings you to $550 million ARR. But because you've got more efficient ASICs online, you're about a 60% margin. So that's $320 million of cash flow based on current mining economics. So just a snapshot for the analysts and all the enthusiast investors out there that like to build their Excel models. Next slide, please.

So we talk a lot about the ROI principle. And just such an avid math and physics student in school, sometimes you come across, you're deep in the chapter, in your physics textbook and there's just this one diagram that just puts the concept visually and then it crystallizes, meaning you have mastery of it and of course, all the equations that backstop. But this is especially for institutional investors that are trying to wrap their head around and they think that it's sort of this elusive or arcane business, it's actually not. Bitcoin mining is the intersect of computer science and electrical engineering is really what it is. We're building high-voltage, high energy infrastructure. And the revenue, the top line is really -- you can almost think of it as combinatorial mathematics. You've got a number of hashes per block is difficulty times 2 to the power of 32. Now Bitcoin price fluctuates, difficulty fluctuates, sometimes difficulty goes down. We just had difficulty dropped by almost 5%. We're 150, 160, now we're at 152. So what does that mean? Well, your top line here, as you see is like a squiggly line. It goes up and down, but with time, and we're talking a 4-year horizon. It does trend down as more hash price comes online. But during the life cycle of an ASIC, we always aspire to get 1 year ROI after electrical costs. And you could see that as in the blue area of the chart. Now once you've realized an ROI on your ASIC, then you're free cash flowing. And so that's why for the balance of that 3 to 4 years, you're free cash flowing. And this is how we mine Bitcoin, and this is how you make money in this business. You have to think this way and you also have to model this way. And of course, you have to invest and operate this way.

Now the 3 bullet points here is, what are the 2 levers on this graph? Well, there's one vertical axis and that's your dollar per terahash ASIC acquisition price. The more you pay for your ASICs on a dollar per terahash basis -- and by the way, this is on a snapshot assuming the certain efficiency of joules per terahash, right? So given a model of ASIC, so a new ASIC comes out, you go great. This is what its ROI profile is going to look like. Depending on what you pay for that ASIC, the more you pay, the longer it's going to take you to hit that ROI. And as there's -- and we know that there's diminishing returns. This is modeled. It's all part of the Bitcoin white paper as more hash rate comes online, it makes the network more secure. Bitcoin is the world's decentralized energy backed currency. So it's not a bad thing that you have to upgrade, and that more energy comes online. It makes the network more secure.

As Bitcoin is a market capital over $2 trillion, you have over 30 gigawatts of compute backing that. And it's the increasing efficiency of the ASICs that helps dollar-denominated to dollar per kilowatt hours, okay? So on the horizon -- on the time horizon, you see the plus/minus dollar per kilowatt hour. You see the lower your dollar per kilowatt hour operating cost, the longer you can run that same ASIC for. So you see how we've indicated or illustrated that additional tail end of free cash flow, well, the lower your power price, the longer you're going to be able to realize free cash flow. And of course, the cheaper your dollar per terahash purchase price, the cheaper that number is. The shorter your ROI duration is, and of course, that means you have all other things being equal, more time to free cash flow.

And this is so important, and it's been moreover, as I said earlier in the presentation, we run our ASICs for their entire economic life cycle. And what that might mean in the third and fourth year, even the second year, on a margin basis, if you're looking at your MD&A, say, what was your gross mining margin this quarter? Well, if you're running older ASICs that are in pure free cash flow scenario, the margin apparently may not seem that great, but they're free cash flow and they paid themselves off. And so you want to run those ASICs until the very end of the life cycle before you upgrade them. And this is an important principle. I thought this visual would very much help. So there you go, Bitcoin mining summarized. Let's go to the next slide.

By the numbers, this is what it translates to. We lead the sector in cash return on invested capital. Quarter in, quarter out, here it is, we've realized 18% annualized ROIC this quarter, 4.5% for the quarter. So I love explaining how the business works. We love to share expertise. We love to be beacons of truth in the industry where there's a lot of hubris. And again, HIVE has been standing strong as the longest crypto miner in the sector for the publicly traded peers, 2 Bitcoin halvings and an Ethereum merge, no one else has navigated all 3 of those events. In addition to that, we built our own ASIC miner with Intel. So we understand the industry at its very depths and also at a very macro level, having strategic oversight with Frank as a money manager, our Executive Chairman and myself as an electrical engineer, we're both numbers' guys, and we like to look at the prism of Bitcoin mining and data centers. So many different angles to understand what does it look like when you evaluate it. And ultimately, our bottom line, though, is how do we get the best ROIC. Here it is. Next slide.

We also have the lowest G&A in the industry. We have a high-performance work culture, our executive team. We were in Chicago for a Jim Collins seminar, really, how do we live and breathe being a good to great business. And so again, high-performance culture. We are in 9 time zones, 5 languages. And again, we talked about a high-performance culture. We are a high-performance culture, bringing 300 megawatts online in 6 months, almost an exahash per week. I think we might have broken an industry record. And so you can clearly see here under 10% of our total revenue is reflected as G&A and -- I mean a lot of our peers are really in the 20% to 30% range is really what it tells you. And so we're able to scale. We have a dual-engine strategy. This is the G&A for the whole corporation. So we're talking all our operating subsidiaries, our global business, HPC and Bitcoin mining, so incredibly privileging.

We work long hours though. So if it's hard to get a call with us, it's -- we'll make time for you, but priorities. Next slide.

So the other amazing thing is the value proposition. So our relative valuation signals a very strong opportunity. On a relative basis, where you look at our peers on EV to exahash basis, we should be at about a $4 billion enterprise value. Now of course, people would say, "Oh, yes, but a lot of your peers here have an [ AR HPC ] strategy." Yes, as do we. And I think we have one of the best AI and HPC strategies with a lot of upside to be realized in 2026 as we bring online those 6,000 Blackwells and pursue hyperscaler colo through New Brunswick as well.

So I think that we also have best uptime in the sector, lowest G&A per Bitcoin mine, best ROIC and right now, the best value. So I think for the smart money, it's an opportune time. We've seen HIVE really rally to almost USD 1.7 billion market cap. There's been a bit of a market-wide correction recently. So I think it's an opportune time as the market consolidates. Next slide.

Now a little bit more about what we do. Yes, we are very high performance, and Frank and I were boots on the ground. We go to all the conferences, but we also invest in the communities. So we had a program where we wanted to refresh 18 schools in the rural regions by our data centers. And so Frank and I were actually visiting in Valenzuela, this is a primary school, elementary school and the kids were so happy. And so the pictures on the left are before, and you can see the beautiful hand painted map of Paraguay. But you see the walls are sort of weathered and chipped. And so what we did was we refreshed the schools. You see brand-new paint and air conditioners. You see all the air conditioners after bolted on the walls for the classrooms and then we installed the playground as well for the kid. We're doing upgrades to 18 schools throughout the community. So just about how we like to invest in communities. As you know, we sponsor the hockey team in Boden, but we do what the community wants and needs. And of course, in this scenario in rural Paraguay, it's the schools. Next slide.

Also, embracing tradition. We were invited to the presidential palace for YPO event. This is Santiago Peña, who is an incredibly gracious host. And so we're wearing a traditional -- the scarfs, sort of traditional Paraguayan dress, and it was absolutely a phenomenal event. We had our whole executive team down there. Of course, Gabriel Lamas and Frank, are pictured here with Peña, Luke Rossy, our COO; and myself and Jaime Perez, our Latin American correspondent. So investing in the communities at the schools in the neighborhoods, but also building strong ties with the leaders of the country. Next slide, please.

Now Paraguay in turn is also building ties globally, and Peña has, I think done a phenomenal job having a great relationship with Trump, all smiles here with the big thumbs up. And moreover, Frank and I had a strategy session with Peña in September during that YPO Summit. And of course, shortly after we were granted an additional 100 megawatts, which we'd announced to the market. So it's about having relationships strategically and being aligned with the government and utility companies and just doing good business. And so we're very bullish on the future of Paraguay, not only for Bitcoin mining, but data centers broadly. Let's go to the next slide.

Directionally, Latin America is emerging with HPC on the horizon, was in the headlines in October, OpenAI and Argentinian energy company are looking at a Stargate for Latin America. So of course, Argentina is adjacent to Paraguay, Argentina buys power from the Itaipu Dam, which is the largest hydro dam by production in the world. And so we believe that there will be a bright potential future for HPC in Latin America. Let's go to the next slide, which is Darcy Daubaras, the longest-standing CFO in the Bitcoin mining industry, and he's going to give you the overview of the financial results. And not only in the Bitcoin mining industry, of course, now we are a HPC business with a thriving AI cloud growth strategy as well as hyperscaler colo for 2026. Darcy, over to you.

Thank you. Good morning, everyone. And thank you for joining us today. And I will be walking through the Q2 results for the 3- and 6-month period ended September 30, 2025. We are providing certain non-GAAP measures in our presentation today. The company believes that these measures, while not a substitute for measures of performance prepared in accordance with U.S. GAAP, provide investors an improved ability to evaluate the underlying performance of HIVE. These measures do not have any standardized meaning prescribed under U.S. GAAP and therefore, may not be comparable to other issuers. Further details are found in the management discussion and analysis for the 3- and 6-month periods ended September 30, 2025. Moving on to the first slide.

HIVE ended September 30, 2025 quarter with 236.8 million common shares, 2.7 million options, 11.4 million RSUs and 5.2 million warrants outstanding. And on the next slide, let's start with the key highlights for the quarter.

For Q2, we generated $87.3 million in revenue, driven by digital currency mining and high-performance computing services and delivered $31.5 million in adjusted EBITDA. Production for the quarter was 719 Bitcoin equivalent, which is up from 406 in the prior period, supported by stable operations, strong uptime across our sites and the execution of our Paraguayan expansion. These numbers reflect disciplined cost management, a focus on efficiency and the benefit of our diverse global footprint. Let's now on the next slide, take a look at how this operational performance translates into our balance sheet strength.

We do take great pride in maintaining a healthy balance sheet. As of September 30, we held $22.6 million in cash, $24.4 million in digital securities and $17 million in receivables and prepaids. That totals $136.7 million in digital assets against $55.5 million in current liabilities.

We also maintained $25.7 million in strategic investments. This strong liquidity has allowed us to manage market cycles, invest in expansion opportunities, specifically our HPC and Paraguayan opportunities and avoiding over-leveraging the company. With that context, let's look at how our earning metrics have evolved starting on the next slide.

Shifting our focus to our gross operating margin on a year-over-year basis, comparing the results of this quarter to Q2 last year, our gross operating margin, which is calculated as total revenues minus direct operating and maintenance costs and high-performance computing service fees increased to $42.4 million in the most recent quarter compared to $400,000 in Q2 of last year. In this most recently completed quarter, we are reporting a basic loss of $0.07 per share compared to a net income of 0, basically flat per share reported for Q2 last year.

Taking a look at our revenue increases year-over-year on the next slide, we generated total revenue in the second quarter of fiscal 2026 of $87.3 million versus $22.6 million in the previous year's second quarter. The revenues compared to the same quarter in fiscal 2025 can primarily be attributed to the expanded hash rate that we are experiencing from the Paraguay expansion. As mentioned previously, our gross mining margin, which equates to our revenues minus direct operating and maintenance costs and high performance computing service fees increased to $42.4 million or 49% margin in the most recent quarter compared to $400,000 or only 2% margin in the prior year comparable quarter. As we've said, that's a direct result of optimizing our mining fleet, the mass of Paraguayan expansion that we've experienced since the beginning of calendar 2025 and improving our overall operational efficiency.

Now if we zoom into just the last 2 quarters, you'll see impressive improvements on the next slide. Comparing our current fiscal Q2 quarter to the previous Q1 quarter, we generated revenue in this current quarter of fiscal 2026 of $87.3 million versus $45.6 million in the previous quarter. The increase in revenues versus the prior quarter was impacted by increased exahash capacity with the Paraguay expansion, an increase in the price of Bitcoin, resulting in higher revenue from digital currency mining and our continued focus on our high-performance computing revenues.

Our gross operating margin increased to $42.4 million or 49% operating margin in the most recent quarter compared to $15.8 million or 35% margin in the prior quarter's comparative. The increase in gross operating margin versus the prior quarter was greatly due to the comparative BTC pricing and resulting revenues and the continued improvement in our fleet efficiency.

And on the next slide, I'd like to remind our stakeholders that our net income is comprised of our operational earnings or cash flow plus our investment earnings, which includes realized and unrealized earnings, which often includes noncash charges. Our adjusted EBITDA in this quarter ended September 30, 2025, was $31.5 million versus adjusted EBITDA of $12 million in the September 30, 2024 period.

I will highlight that adjusted EBITDA is a non-GAAP figure. For this completed quarter, we experienced loss of $15.8 million compared to a net income of 0 in the previous year comparative. The net loss experienced this quarter was driven by noncash losses experienced on our strategic investments and changes in the fair value of our derivatives, which is linked to the Bitcoin used as deposits on equipment purchases.

On the next slide, the quarter-over-quarter view tells a similar story. Our adjusted EBITDA in the second quarter of fiscal 2026 was a profit of $31.5 million versus adjusted EBITDA of $44.6 million in the previous 2026 Q1 quarter.

In the second quarter of fiscal 2026, we experienced net loss of $15.8 million compared to a net income of $35.5 million (sic) [ $35.0 million ] in the previous Q1 quarter. The Q2 fiscal 2026 was a solid quarter for HIVE. We delivered strong revenue, expanded margins and maintained a robust balance sheet.

Our operational discipline, fleet expansion and cost control measures continue to position us well to compete in a challenging environment and capture opportunities for growth. As always, I want to thank our loyal stakeholders and encourage them to continue to follow our expansion efforts in both Bitcoin mining and high-performance computing operations. There's exciting things to come, so stay tuned.

Thank you, Darcy. That concludes the presentation for today. We will now begin the question-and-answer portion of our call. [Operator Instructions] Our first question comes from the line of Darren Aftahi from ROTH.

Congrats on the progress. Two if I may. Just can you talk about what the assumed spend on CapEx per megawatt is to either retrofit or greenfield your data centers to Tier 3? And then on the AI cloud strategy, I think, Aydin, you talked a little bit about financing and then leasing the GPUs. I guess, do you guys plan to buy these at the end of that cycle? And then I guess, what's the assumption on useful life you're assuming on GPUs?

Yes, you bet. So the site in Boden is $25 million to build out to convert and Toronto is $40 million. They -- we've had those facilities effectively quoted and we're ready to commence the conversion of those. So the price per megawatt is different because each facility has just different existing infrastructure. Boden has already designed as GPU facility, and that's the one that you've been to actually when we did the site tour last year. And Toronto is the facility that will -- that we recently announced a purchase of. So again, that's $25 million and $40 million, respectively. And then that provides -- and then, of course, Bell is colo. So there's no significant CapEx. So that collectively gets us to the footprint for 6,000 GPUs. And then -- sorry, what was your second question?

Really just on the cloud...

The buyout. It's really -- it's like really a lease-to-own. So the buyout is de minimis. So we will effectively have at the end of the term for a very nominal amount ownership of the GPUs.

And what's the useful life you're assuming on this? I know you might be one of the fewer entities that have...

I think broadly, what we've seen is that after 3 years, the -- easily or conservatively, GPUs are worth half of what you bought them for. And then I think that you could see a useful life conservatively is also 5 years. And I back that up with a couple of data points. As you know, we've -- as you've been following the story for a long time, we bought 38,000 Ampere GPUs, A40s, A6000s, A5000s, A4000s. And we made that order. We announced in July of 2021. So a lot of those GPUs to the tune of over 4,000 are in our AI cloud still running. We have 4,300 A-Series cards, mostly A40s and A5000s. And here we are 4 years later, and they're still running. So that's why I put the useful life at 5 years. I believe it could be longer. I believe it could be 5, possibly 6, maybe even 7.

And then on the resale, we sold, and this is sort of in our past disclosures and M&A, we've sold the balance of that fleet of Ampere Series GPUs, about 34,000 of them at approximately 90% of their original face value. And those -- the sale of those GPUs was last year over kind of the course of the year. And so that puts us about 3 years into the economic life cycle, and we were selling them for 90% of what we bought them for.

So again, we're very strategic about ROIC, and we realized that there was a great opportunity there to get such a high value for them. And we took those proceeds and then we expanded instead and got more Hopper H100 and H200 GPUs. And that get us to where we are today. But moreover, to answer your question, we were able to sell at 90% of face value after 3 years. Of course, that was driven by global demand. And so when I told you 50%, I'm trying to give you very conservative, realizable figures.

So to wrap it up 50% after 3 years, and I think a useful economic life cycle of 5-plus years. And look, that 3 years, it could be higher, maybe the 2Ps are worth 70% more. But I mean, in our case, we've seen 90%, but let's just say, 50% to be conservative after 3 years.

Next question coming from the line of Mike Grondahl from Northland.

My question, on those 6,000 GPUs that you're putting out the next couple of quarters, how many of those are contracted today? Meaning what kind of visibility do you have on those? And then could you talk a little bit about the average term, at one point, I think I heard 3 months to 12 months and then another 2 years and just kind of the demand you're seeing?

Yes, Mike. So the -- you're correct in that the existing GPUs that we have online are in the 3- to 12-month range. And the -- I actually just want to refer you to -- sorry, I'm in Paraguay right now, someone's knocking on the door. So we also put out a press release this morning concurrent with our Q2 earnings that actually announced the order of the first cluster of 500 for those Blackwell B200s. And so that's the first cluster that we expect to be delivered before Christmas this year.

And so the 6,000 GPUs, there's a chart in my section where we actually saw the first 1,000 of the 6,000 coming online by Q1, 2026 calendar Q1. The second cluster of 1,000 GPUs comes online Q3 2026. And then finally, the additional 4,000 Blackwells would be Q4. And the reason why Q4 is, again, it's the retrofit of the Boden and Toronto facilities. And those retrofits are underway. So of course, you wouldn't be bringing GPUs online until that conversion was complete. And so you've got 1,000 in Q1, another 1,000 in Q3, another 4,000 in Q4.

So to answer your question in short, those GPUs are not live yet. We're bringing them online in that cadence. And so we will provide more color on the contracted GPUs when they go live. So I hope that answers your question.

And why 2 years is because we wanted to provide sort of a conservative base case if we underwrote the revenue of the GPUs on the longest-term contract, which conversely provides the lowest dollar per GPU hour. And so we're aware directionally that the Street and institutions are more in favor of the longer the cash flow, even if it's a lower quantum on GPU per hour, people like the stability. And so we've modeled that. But then moreover, with the cloud, what's exciting is when you're selling tokens or allowing people to utilize AI enterprise services with API endpoints, now you're selling a much higher margin. And so that's why we kind of had 2 similar slides showing the growth in revenue. One is on long-term 2-year contracts and the other is higher margin selling AI enterprise services. So I believe my comment -- I hope my comment on those slides made it clear that we would expect the realized ARR to be sort of in the midpoint somewhere between those 2 projections I provided. Does that help?

Next question coming from the line of Bill Papanastasiou from [ Stifel ].

Congrats on all the progress made scaling the company's Bitcoin mining and AI, HPC initiatives. First one on Bitcoin mining. Are you able to share your thoughts on where you think hash rate and economics could trend following the industry's push to AI, HPC and just general commentary on the strategy going ahead?

Yes, for sure. Thanks, Bill. So we currently are at a difficulty of 152 trillion difficulty downward corrected recently after 156 trillion all-time lined -- 156 trillion being an all-time high. Now what I would point out is the last correction of difficulty going from 156 trillion to 152 trillion, we actually were at a -- before the adjustment, we were sort of at a hash price in the 42 -- now I'm going to explain this sequentially because there's 2 variables, Bitcoin price and difficulty. So right now, I'm referring to when difficulty adjusted. And of course, Bitcoin price was in the $102,000 range at that time. So a Bitcoin price at the $102,000 range at that time, we saw hash price sort of in the low $40 per petahash per day range, and that downward adjustment difficulty brought us up a bit to the -- kind of going from $41 to $42, up to $43 to $44. Obviously, as difficulty drops, hash price increases.

So we sort of saw that. But then recently in the past sort of 5 days here, we've seen Bitcoin drop from $102,000 to $99, and now we're sitting at $93.8. So that's put hash price back down to $40 petahash a day. So I think consistently, what we've seen since the halving is that the hash price floor on a 7-day moving average has been $40. Now I guess, actually at $93,000 a $39 hash price. So it will be telling to see what happens over the next epoch. The next difficulty adjustment will be November 26. So we're still about 10 days away from that.

And if the hash price floor of the last year is an indication, we would expect a downward adjustment, although as more efficient machines have come online over the course of the last year since the halving was in April of 2024. And we're now in 2025, we could see the hash price floor perhaps become incrementally lower, perhaps in the $36 to $38 range. But we'll know sort of in the course of next week. A lot of the peers are not publicly -- when I say peers, publicly traded Bitcoin miners, have not announced significant expansions. Of course, Bitdeers is being a producer has brought online more hash rate. I think Cipher has got themselves to 23 exahash. MARA, IREN, they're all hovering about 50 along with CleanSpark. So that hash rate is online. I don't see that going anywhere anytime soon. However, again, we did see difficulty downward adjust. So there will certainly be smaller-scale miners perhaps with less efficient machines that are shutting off their hash rate capacity when we see hash rate in the $40 -- $39, $40 range.

So we're definitely closely watching it. But the good thing is our breakeven hash price is about $22 petahash a day on a global fleet adjusted basis with a 17.5 joules per terahash global fleet efficiency. So we've mined every quarter since I've been with HIVE 2021 profitably. And so we upgraded our fleet accordingly, very intentionally to be able to mine with positive gross mining margin through the next halving.

So look, with the Fed coming out, hopefully, there's a rate cut in December and that quantitative easing cycle. Hopefully, there's more liquidity, and we see a rally in risk on assets such as Bitcoin. Time will tell, but we're very much aware and monitoring things, but in the meantime, continuing to mine profitably.

Appreciate that color, Aydin. And then another one, if I may. Within the HPC stream, curious how we should just be framing our estimates between that base case and blue-sky opportunity? Are you able to share your targeted composition in...

Yes, yes, that's a good question and a very good question. I think as I sort of alluded to with Mike, we expect it to land somewhere in between the base case and the case with the additional -- or I should say, higher margin revenue with the cloud stream -- sorry, with the Buzz platform stream enabling higher margins on the cloud business. But I think you'll have your answer sooner rather than later because as we -- again, we're bringing these GPU clusters online first with Bell. So again, we expect those 1,000 GPUs to come online through the course of Q1. Again, this morning, we announced the purchase of the first 500, which we expect to get delivered before Christmas.

Now, of course, takes time to plug and configure and pressure test these clusters. It's -- Frank often likes to give a very helpful analogies. From 6 hours of receiving an ASIC, you could have it hashing. It's more like 6 weeks when you're building an InfiniBand cluster with a bunch of GPUs. So we're -- we've got our first cluster of GPUs live in January, this 500 January, the second 500 coming online within a month or 2 of there, you'll see sort of what those figures are. So effectively, just stay tuned for fiscal Q1 -- sorry, calendar Q1, which would be fiscal Q3 period end March, and you'll see those numbers.

Next, we'll go to the line of Chris Brendler from Rosenblatt.

Congrats on the results, very impressive progress here. I just want to ask a question on the Bitcoin mining business for a second. The improvement in gross margins here, and I know the network has gotten more difficult, more challenging from a competition perspective, but your gross margins improved significantly quarter-over-quarter. And with the sort of a full quarter run rate of the Phase 3 complete now, I would expect the gross margins to continue to improve. Just wanted to get your -- take your temperature on that thought that your hash costs are going to continue to decline and your gross margins, all else being equal and the Bitcoin mining business should continue to improve now that you're fully up and running on Phase 3.

Yes. You've got it bang on, Chris. So the Q2 -- fiscal Q2 we just reported, the average cash rate for that quarter was 16.1 exahash, and that was a period end September. And so that hash rate, I believe that it was Phase 2 of our Paraguay project. So 200 to 300-megawatts was completed in that quarter. However, today, we're at 25 exahash and that milestone, we just reached last week. And we had sort of provided an update on that, that we put out a press release that we hit 300 megawatts -- we completed 300 megawatts, which is super exciting. We hit 24 exahash operational, 25 exahash were installed. And there was some short-term just seasonal curtailments we're talking like a day or 2 at a time here in Canada. And then just a few days ago, we actually hit the full 25 exahash operational as well. So that's super exciting. But of course, we're in November now.

So to your point and to further answer your question, yes, more hash rate coming online with our latest generation, 15 joule per terahash, S21+ Hydro machines are hitting the tape, and therefore, it will bring that average hash cost down as well as our cost of electricity in Paraguay is cheaper. And so you're getting 2 factors there that are attributing to improved gross mining margins. So yes, that's right.

And so again, just for your model there, I believe it was 16. 8 exahash was the quarterly average operational hash rate for a period end September. Again, that's covering the growth from July, August, September. We obviously ended the period at a higher number. I think it was closer to 20, but nevertheless, you had quite a ramp. So what's the average of that ramp for the entire quarter? I believe it was 16.8. And until this quarter, it will not -- you could sort of interpolate the data just simply using our monthly production reports, what was October, which we put out not too long ago, and you'll see November reports in December. So you'll be able to figure our average operational hash rate. And again, all the new hash rate we brought online is at 15 joules a terahash. The global average with the full 25 exahash online at 17.5 joules a terahash. And so we believe the monthly production reports are helpful for the analysts when doing your model. And we know some of our peers have sort of stopped doing monthly production reports.

I don't understand -- I mean, I know why. Broadly, they're sort of trying to rebrand themselves or perhaps not advertise the fact that 99%, 97% of the revenue still comes from Bitcoin mining. No names. I mean, everybody knows. But yes, we -- I believe Bitcoin is the world's energy back decentralized currency, and we've had wonderful year transforming our business to institutional scale as that business is flourished. I believe pound for pound, we are the best Bitcoin miner in the sector. We lead in ROIC as evidenced in our quarterly metrics again. And so it's not something we're going to shy away from. Of course, HPC business is in Buzz, which is our pure-play subsidiary. So -- but it's part of our dual engine strategy. So there you go. I hope that's helpful.

Yes, sure. My follow-up question is on capital strategy. I mean at a high level, just the tremendous amount of growth you've had over the last year, and you've got big plans for next year as well between HPC and Bitcoin mining. Phase 4, I guess, is now on the table for Paraguay. How do you think about balancing the capital needs between HPC and Bitcoin mining? Is there a preference at this point? Or are you still looking to grow both at sort of the same rate?

So in my presentation, what I outlined was we have the additional 100 megawatts in Paraguay, which was awarded to us. So currently, our global operating capacity is 440 megawatts, with 300 megawatts in Paraguay. Now we have been granted an additional 100 megawatts of green energy by ANDE in Paraguay, which is great. What's attractive about that is that it's actually at an existing site, the Yguazú site when we bought, that was actually designed and all the civil work has been done for 300 megawatts at that site alone, but we've built it as a 200-megawatt site because those are the transformers that were ordered and that's sort of what Bitfarms had near term. That's all being completed.

So to get that extra 100 megawatts online, all we're doing is bringing on a couple more transformers for the substation and the medium voltage and then dropping hydro mining infrastructure in place. So sum that all up, the punchline is $0.25 million a megawatt to expand that 100 megawatts, which is incredibly attractive as we've mentioned before, our all-in cost for land, substation, everything was $400,000 a megawatt for our Valenzuela site and our Yguazú site. But because this is an expansion of an existing site that's already complete, it's a nominal incremental cost.

So that's one portion of the CapEx. The bigger portion, of course, is I think all the analysts on the call and a lot of the very enthusiastic retail investors and certainly, institutions are aware that a cluster of Blackwells like a 1,000 cluster with InfiniBand and everything is roughly $60 million. So if you're bringing on 6,000 GPUs, about $360 million, but that's a significant CapEx. But if you've got vendor financing for that, then that is the most significant component of CapEx that you've now alleviated. So we -- at least in terms of relying on equity financing or, say, taking on debt in the capital markets or having a derivative like a convert or anything. So what we're really enthusiastic about is that with vendor financing, we are able to scale that GPU business. And again, Bell being a colo, there's no significant CapEx for us. So that ramp to the first 2,000 GPUs of our projection, getting us from $20 million to $60 million of ARR, we're able to realize through the colo with Bell and vendor financing on the GPUs, no significant CapEx, which is really nice.

And then, of course, as I think it was Darren's question, the only CapEx we have to finish is with Little Boden and Toronto, $25 million and $40 million, respectively. And that gives us our ramp in our pipeline to operate 6,000 Blackwells globally in addition to our existing fleet of 5,000 GPUs to the tune of 11,000 GPU cloud globally. So yes, it's about being very balanced and strategic with how we deploy capital and where we can have what we believe is accretive leverage, then we'll pursue that. And we have, and I'm very happy to announce that we've now got this capital plan laid out in addition to really support the growth.

Time for a few more questions, 2 more. Next question comes from the line of Fedor Shabalin from B. Riley.

Thank you very much for putting out monthly updates. That's helpful. And question-wise, could you please provide an update on the time line for the Brunswick data center, including whether construction will be phased or completed at once? And additionally, should the colocation agreement not proceed hypothetically, is there a path to fully utilize the capacity for AI cloud demand? And how long would it take in this case, including timing for arranging, financing for GPU procurement?

Yes. So I think the takeaway from the presentation we provided, we actually wanted to emphasize that the conversion of the New Brunswick site, especially given its proximity close to the border of Maine, actually makes it very opportune as a hyperscaler colo project. And we're very much aware that right now, especially institutions, we were in Miami for the Cantor Conference, which was an absolute home run. I think Cantor did a phenomenal job putting together that event was extremely well attended, the quality of the attendees on both the buy side. And I mean, you basically had the whole industry and all the Wall Street there. But anyway, it's clear that the Street is valuing hyperscaler colo because they like the stability, and they like to scale of those revenues. If you want to do a 100 or 200 or 300-megawatt site for hyperscaler colo, you can now allocate a large amount of capital. And if you've got a long-term leases, we've seen some of our peers announce, now you've got stable cash flows at lower margin.

I mean just -- I think all the analysts on this call are well aware, but maybe for the other people that are dialed in, just if you look at it arithmetically, if you were to sign a lease, one of our peers recently did a lease that worked out to $102 per kilowatt a month on hyperscaler colo, all you do is you divide that by 730 hours, and that works out to $0.14 a kilowatt hour. So the unit economics of hyperscaler colo are relatively modest, $0.14 a kilowatt hour when Bitcoin was at a $42 hash price, ASICs were doing $0.12 a kilowatt hour, not too far off from that. $55 hash price. Our ASICs were doing $0.15 a kilowatt hour, that's 21 plus. And you're base operating costs in mining is with power and so forth about a $0.05. So you're doing relatively high margins and quicker ROI. If you're a $50 hash price range, something like 1.5 years ROI after CapEx.

Now that's why I like the dual engine strategy. But again, because we're aware that people like the scale and stability of hyperscaler colo, even though the ROI period is more like 7, 8 years, depending on what your dollar per megawatt CapEx is to do the conversion with a long-term contract. It is being valued in the Street. So right now, as we've positioned our 2026 targets, we've got both. We've got the revenue, right, from operating AI cloud, hyperscaler colo as we've targeted for New Brunswick.

And then, of course, the Bitcoin mining business. So the $85 million target, Fedor, was actually based on hyperscaler colo. And so we did want to point out though that if you were going to base it off of GPU reference architecture based on latest generation GPUs, you could operate 25,000. But right now, our target and our focus is to get through the conversion of Toronto as well as Sweden and the Bell partnership to bring online those 6,000 GPUs. And stay tuned for updates on the strategy and status of the conversion of New Brunswick and directionally which way it goes.

But again, I sort of emphasize that the projections we provided in both cases, I'd had those 2 projections on HPC revenue. Both cases were predicated with New Brunswick as hyperscaler colo. So I would say just stay tuned for updates on that strategy and developments there.

Final question comes from the line of Joe Vafi from Canaccord.

Terrific progress. Great to see it. Just can we drill down a little bit on Itaipu and Paraguay, clearly becoming very strategic location for HIVE and very differentiated. And I know you did mention the Stargate initiative in Argentina, Patagonia. Clearly, Stargate like the clean power angle as well. You kind of just drill down. I know you've been down there. I know Frank has been down there a lot. You're close with the government and the President. How much power do you think -- is there additional power available to HIVE out of Itaipu? And maybe what's the President of Paraguay thinking about AI and things like that as the industry develops?

So what we've noticed is the -- I mean, the President of Columbia -- sorry, the President of Paraguay, Santiago Peña. He's like -- the ministers, the President, they're all like internationally educated. They're all extremely bright and the -- I mean Peña, I think he's 46. I believe it was his birthday, he just turned 47. So these are young, really smart guys, Joe, that want to position their country. And they've been internationally educated, I mean, in the U.S., I believe. Peña was studied at Columbia University. So they want to position their country for success, and they're aware of they're very much attuned of -- they're very much attuned of the data center industry.

We saw that OpenAI went down to Argentina to talk about a Latin American Stargate with Sur Energy. We, of course, have been talking with the Paraguayan government about the future of data centers in Latin America and HPC in Paraguay. So I think that it's definitely on the horizon, and it's something that we're incubating and seeing how we could support and be a catalyst to help bring that vision to life. I would say that there's a vision for it right now and where -- what we're trying to do is be a catalyst for that vision in Latin America by working very closely with Peña and the various ministers, again, who are extremely business savvy. And Paraguay is a country of 6 million people, but it's also home to the largest hydro dam by production.

The Itaipu Dam, by the way, maybe I'll put it on Twitter, Luke and I took a photo. It has again this world record banner when you approach it. So it actually kind of set up as a tourist attraction like anyone can go there, marvel of engineering, et cetera. And so how do you monetize those vast natural resources? And I think Frank has often talked about how as we grow our footprint going from 300 to 400 megawatts in Paraguay. That means U.S. dollars paid in a timely manner for the government of Paraguay. And moreover, we believe that directionally, power prices are going to get attractive in this region. And they are attractive, and I think they're going to more so become attractive. And -- that's sort of what I could say for now. But let's just say we're working on a lot of things and standby for updates as we have them.

That's great. And then just quickly on a follow-up. I think the Brazilians and the Paraguayan share the power off that dam. I think it's -- I mean, a Wikipedia search, I think it's like 15 gigawatts coming off of Itaipu. Have you had any discussions with the Brazilians at this point? Or -- because I think they're probably getting a lot of the power off the dam. Just wondering if you're exploring other things in the region given your Itaipu position.

Yes, you're very astute, Joe. So yes, it's about 14 gigawatts. When we were there in December of last year, it was producing 10.7 gigawatts on that day. And we actually -- Jamie Brown came down for a visit in September. There was a YPO event that Peña hosted. And it was at the Central Bank and then it was at the Presidential Palace and then we went over to Itaipu. And I believe that the dam was only producing about 8 gigawatts that day or -- yes, exactly. So that dam has a lot of capacity, exactly 14 gigawatts. Might be 14-point something, closer to 15. But when we were there, I recall it being 14 gigawatts.

So there is -- the arrangement is that the power is actually sold like while -- when you go to the dam, it's called Itaipu Binacional. So there's a border and you've got both Brazil and Paraguay, but the power is actually sold to Brazil. So I think it's -- what happens is there's sort of a more nuanced outlay where the Itaipu infrastructure project has recently completed its CapEx, the construction cost of this magnificent facility, it has paid itself off. It has paid itself off. And so now what we're entering to in the 2027 era is what will that look like going forward now that it's been paid off. And so when I sort of was referring to, we believe that the pricing of power from the Itaipu Dam will directionally be favorable. That's what I was sort of getting at. So you've asked a very good question. You've clearly done your homework. But that's all I can say for now, my friend. And stand by for updates on that.

But yes, you're kind of pointing in the right direction. And again, we've -- with OpenAI coming down to Argentina, of course, Argentina is a neighbor of Paraguay as well. We think that there's going to be some very exciting potential activity, both in terms of having abundant green energy. And what we're also doing is initiatives to help improve the grid as well. And so part of -- as we showed how we were refurbishing the schools in the community, and that's goodwill at HIVE, right? No one's asked us to do that. We've done that because we think it's the right thing. And Frank and myself and our whole executive team, we really much value education even if it's at a primary school level. Of course, we have great education for the staff that work at our facilities. Frank is a huge sports guy as anyone who knows him knows. And so we love to sponsor sports teams as well.

But then you kind of take it up a level and when you're working with the utilities and the federal government, how can we help improve the grid and contribute to the nation's infrastructure. So standby. But yes, we're working on things that we'll provide you updates on over the course of the next year.

Excellent. Thank you, Joe. Thank you to all of our analysts and our attendees. We appreciate you joining us this morning. That concludes our Q&A session in our Q2 2026 earnings call. Thank you for joining. Thank you for your support. We look forward to speaking to you again soon.