WhiteFiber Stock

Hello, and welcome to the White Fiber Second Quarter 2026 Earnings Conference Call. Good morning, and thank you for joining us. We will begin with prepared remarks from manage. [Operator Instructions] As a reminder, today's conference is being recorded. I would now like to turn the call over to your host, Cameron Schnier, Senior Vice President of Capital Markets and Corporate Strategy at White Fiber. Cameron, please go ahead.

Thank you, and welcome to the White Fiber Second Quarter 2026 Earnings Call. Joining me today are Samir Tabar, our Chief Executive Officer; and Justin Zhu, our Chief Financial Officer.

Before we begin, I'd like to remind everyone that some of the statements we make on this call are forward-looking in nature and subject to risks and uncertainties that could cause actual results to differ materially. Such risks and uncertainties include, but are not limited to, those factors described in today's earnings press release, our Form 10-Q for the quarter ended June 30, 2026, filed today as well as other filings we may make with the SEC from time to time. Our remarks today may also include non-GAAP financial measures. Reconciliations to the most directly comparable GAAP measures can be found in our Form 10-Q and in the earnings press release posted on our website. Following our prepared remarks, we will open the line for questions. With that, I'll turn the call over to Sam to discuss our performance. Sam?

Thank you, Cam, and thank you, everyone, for joining us. Last week marked the first anniversary of White Fiber's initial public offering. Over the past year, we've made substantial progress towards the company we set out to build. Most notably, we signed a transformational 10-year agreement, representing approximately $865 million of contracted revenue for 40 megawatts of IT workload at NC1. We've since advanced the project through construction and now into active customer deployment. We've also started operations and turned on revenue at our Montreal 3 location under our Cerebras agreement, expanded our development pipeline, strengthened our capital base and repositioned our cloud services business around larger, longer duration opportunities.

We are proud of what we've accomplished in our first year, but we aren't satisfied. We remain in the early stages of what -- we remain in the early stages of what White Fiber can become. Our first -- our most significant accomplishments remain ahead of us. As we enter this next phase, I'd also like to welcome Justin Zhu as White Fiber's Chief Financial Officer. Justin previously served as Senior Vice President of Finance and Chief Accounting Officer. He has been with White Fiber since its formation. He has a deep understanding of our business, financial operations and growth strategy. Eric is stepping away from his executive role at White Fiber to focus fully on Bit Digital. We thank Eric for his important contributions to White Fiber's development. Eric will continue to support White Fiber as a senior adviser and nonvoting observer to our Board. He will provide additional continuity through the transition. We believe this structure provides each company with increasingly dedicated financial leadership as both businesses continue to grow.

Turning to our operating update. I'll begin with NC1, which remains our most important near-term operating and financial priority. NC1 has moved into active customer deployment. As of today, approximately 20 megawatts of IT capacity is available to support the installation and testing activities of Nscale and its investment-grade offtaker. Initial billing to our customer has now commenced for the initial tranches of capacity. Remaining equipment start-ups and testing are progressing very well. We expect the remaining capacity to be turned over progressively through August. By the end of this month, the full 40 megawatts of contracted IT load will reach a full run rate billing.

As we discussed last quarter, the pace of the ramp was affected by delivering and commissioning issues involving certain switchgear equipment. Those issues have since been resolved. Final deployment also requires tight coordination between the commissioning of our infrastructure and the installation and testing of customer equipment. We worked closely with Nscale on a phased turnover schedule that sequences the work being completed by both parties. While the ramp has taken a touch longer than we originally anticipated, the contracted economics of the agreement remain unchanged. The results speak for themselves.

It took disciplined coordination across our team, our customer, the utility, our equipment vendors and our construction partners, all amid persistent supply chain constraints. We believe NC1 shows what White Fiber can do. It demonstrates our ability to execute complex large-scale AI projects. Just as importantly, we have expanded an experienced operating team on the ground. The team spans facility operations, engineering and customer support. This is not simply a development project or a piece of powered real estate to us. It is a mission-critical facility built to operate continuously and support customers over long-term contracts. The people, systems and operating capabilities now in place reduce execution risk as NC1 moves towards full contracted operations. We established a foundation for continued expansion of the campus. We also established a long-term and positive presence with the local community. Ultimately, we're building a durable operating business in North Carolina.

The initial 40-megawatt deployment is only the first stage at NC1. We expect Duke Energy to provide a delivery schedule for the next 45 megawatts of gross capacity in the near term. At that point, Nscale will receive priority notification of the available capacity in accordance with our existing agreement. We also received extremely strong inbound interest for this new upcoming tranche. We'll evaluate the path forward based on what we believe will deliver the best outcome for White Fiber.

Beyond this, we are working with Duke Energy in further evaluating the potential delivery of an additional 200 megawatts of incremental power to the site. Together with the initial phases, that would bring NC1 to approximately 300 gross megawatts. This is a longer-term opportunity and remains subject to the utility process. We believe it shows how NC1 could scale over time. It also shows why securing the site early was strategically important. NC1 is our flagship facility. It validates White Fiber's ability to acquire, develop and operate large-scale AI infrastructure. We intend to repeat that capability across our pipeline.

Turning to our Canadian portfolio. The most significant update is at NTL2. We had paused development while we evaluated the best use of that site. We've now decided to move forward. We plan to develop approximately 5 megawatts of gross capacity targeting completion around year-end. This decision is supported by active discussions with certain prospective customers. We're evaluating 2 deployment paths. The first is traditional colocation. The second is a vertically integrated deployment combining our data center infrastructure and our cloud services capabilities. We'll provide more details soon as customer discussions and the commercial structure progress.

Moving on to our other sites. MTL1 continues to perform steadily. Recent customer renewals support a stable outlook, and we're evaluating a modest expansion of that particular facility. At MTL 3, the Cerebras deployment continues to perform well. We're also pursuing additional utility capacity for that site that could support a meaningful expansion over time. The approval process remains ongoing.

Beyond our existing portfolio, demand for power-ready, high-density AI infrastructure remains very strong. Demand is particularly acute for 2027 deployments. This reinforces our view that capacity able to reach the market within the next 12 to 18 months will remain extremely scarce. This is where our retrofit-first approach has a clear advantage. We prioritize sites with existing infrastructure and a credible path to power. That allows us to bring capacity to market faster than in traditional greenfield development.

Speed to market is a key competitive advantage for White Fiber. We've built a substantial development pipeline. We're concentrating on the opportunities we can advance towards definitive commitments. We remain disciplined with capital. We prioritize sites with clear current and future power visibility, strong customer alignment, attractive return potential and a path towards project level financing.

We're also deliberate about sequencing our investments. As permanent financing for NC1 progresses, we expect greater flexibility to advance the next opportunities in our pipeline. We remain focused on moving forward on the right terms and in a way that supports disciplined, repeatable growth.

Turning to cloud services. We made substantial progress in transforming the business around larger, longer duration customer engagements and a more capital-efficient operating model. We streamlined the organization and concentrated our resources on the areas where White Fiber provides the greatest value, that being sourcing next-generation hardware, deploying complex clusters and operating infrastructure over the life of a customer engagement. We're encouraged by early results.

Our commercial pipeline has expanded considerably. We are increasing converting that pipeline into larger scale multiyear contracts. These agreements are supported by firm customer commitments. Customer prepayments and third-party equipment financing significantly reduced the equity capital required from White Fiber's balance sheet. We're also seeing an important shift in how customers select infrastructure partners. Larger buyers are consolidating their deployments among a smaller group of providers capable of supporting them at scale. While pricing remains important, customers are increasingly prioritizing engineering credibility, deployment execution and reliable ongoing operations. We believe these are the areas White Fiber is particularly well positioned.

Since our last earnings call, we've entered into new multiyear cloud services agreements representing more than $540 million in aggregate contract value over their initial terms. Based on contracts signed to date, our cloud services portfolio is expected to generate more than $200 million of annualized revenue once fully deployed. One of the new agreement is with Base 10, an AI infrastructure platform focused on production inference workloads. Under the 3-year agreement, we will deploy 1,392 NVIDIA B300 GPUs at a third-party data center in Ontario. The agreement represents approximately $165 million of contract value over its initial term with service targeted to commence in November of this year. Phase 1 also has the option to extend the deployment for up to 2 additional years, creating potential of upside beyond this committed initial term.

Separately, we entered into a 3-year agreement with Prime Intellect, an AI-focused platform on large scale -- focused on large-scale model training and distributed compute. Under the agreement, we'll deploy 576 NVIDIA Ver Rubin 200 GPUs in Canada, marking White Fiber's first Ver Rubin deployment. The agreement represents approximately $108 million of contract value with service targeted to commence in the second quarter of 2027. This Ver Rubin deployment demonstrates the technical depth and expertise of our engineering team. It also aligns to our strategy of focusing on current and next-generation GPUs. Both of these deals expand existing customer relationships, and that illustrates our customers' confidence in White Fiber's engineering and operational capabilities. We also continue to advance our previously announced 5-year deployment in the Paris region, which represents over $160 million of contract value.

Following the completion of procurement and site level arrangements, we're targeting an end of September ready for service date. Additionally, we entered into a 5-year agreement with an existing customer supporting the deployment of 576 NVIDIA V300 GPUs in Iceland. The agreement represents approximately $87.5 million of contract value over its initial term with additional potential upside through revenue sharing. We expect deployment to commence later this year. Beyond these dedicated infrastructure deployments, we're seeing meaningful demand for our managed services offering. Under this model, customers fund the underlying hardware and data center capacity, while White Fiber applies its technical and operating capabilities to deploy and operate the infrastructure on their behalf.

Managed services would allow us to generate revenue without funding the underlying equipment, creating a hyper capital-efficient path to growth. This model will also leverage systems and personnel and expertise that are pretty much already in place. This creates the potential for attractive incremental margins with limited additional direct operating expense. We're in active discussions regarding several potential managed services engagements, including larger scale opportunities. We believe managed services can become an increasingly important capital-light extension of our business.

To support cloud growth in 2027 and beyond, we've entered into an agreement with data center developer and operator, Krambu. The agreement provides White Fiber with exclusive access to 100 megawatts of liquid cooled colocation capacity beginning in 2027 with the potential to expand over time. Access to deployable power remains a key constraint across the industry. This agreement provides an important pathway to additional capacity for our cloud services business.

Taken together, these developments demonstrate the progress we're making toward a scalable cloud services model. We can secure access to deployable capacity. We can provide dedicated infrastructure through long-term customer commitments. We can access third-party equipment financing, and we can apply our technical expertise to customer-funded infrastructure through managed services engagements. These models allow us to pursue longer duration revenue while maintaining discipline around WhiteFiber's capital investment. Finally, we continue to advance our cross data center networking initiatives.

During the quarter, we successfully demonstrated 111.2 terabits per second of bandwidth with guaranteed sub-millisecond latency across 83 kilometers. We believe our patent-pending technology has the potential to create significant platform value for White Fiber. By enabling certain AI workloads to operate across geographically separated facilities, it could allow us to aggregate smaller blocks of power and compute into a single integrated environment, thereby creating a virtual super cluster under one logical system. This could expand the commercial utility of capacity that might otherwise be difficult to monetize independently. This would also increase the value of WhiteFiber's broader site portfolio.

We're now validating specific customer cases for this technology. We're targeting an initial commercial launch of this new technology by this September. Given the proprietary nature of the architecture and the early stage of commercialization, we're not disclosing all aspects of the technology and commercial mode for now. Over time, we believe this opportunity could extend beyond WhiteFiber's own infrastructure through licensing and other commercial structures involving third-party facilities.

Across both colocation and cloud services, the demand backdrop remains extraordinary. We're being deliberate about how we grow. Our priority is to pursue the right sites, customers and deployments. We will scale at a pace that allows us to execute consistently, maintain a high standard of service and continue building White Fiber's reputation as a trusted infrastructure partner. I'll now turn the call over to our Chief Financial Officer, Justin, to discuss our financial results. Go ahead, Justin.

Thanks, Sam. Second quarter revenue was $28.8 million, an increase of 54% from $18.7 million in the second quarter of 2025. Cloud services revenue was $23.8 million compared with $16.6 million in the prior year period. Revenue for the quarter included approximately $12.3 million associated with the previous disclosed customer termination. The termination also result in approximately $4 million of related expenses payable to the GPU lease provider, which was recorded in cost of revenue.

Underlying cloud services results also reflected temporary downturn between the termination of prior contract and the commencement of the newly signed replacement contract. Colocation revenue was $4.7 million compared with $1.7 million in the prior year period. The increase primarily reflects the contribution from MTL 3, which commenced operation under our agreement with Cerebras in fourth quarter 2025. Gross profit, excluding depreciation and amortization was $17.1 million, representing a gross margin of approximately 59% and this compared with gross profit of $11.5 million and gross margin of approximately 61% in the prior year period. G&A expense was about $14.8 million, down from the $17.8 million in the first quarter.

The sequential decline primarily reflected lower professional and consulting expenses and lower share-based compensation expense. G&A for the quarter also included approximately $2.2 million of bad debt expense associated with the previous disclosed customer termination. Adjusted EBITDA was about $5.5 million compared with $3.3 million in the prior year period. A reconciliation of adjusted EBITDA to net loss is included in our earnings release and Form 10-Q. Net loss was $15 million or $0.39 loss per diluted share. The net loss reflects a higher depreciation and interest expense associated with the expansion of our infrastructure and related financing activities. We ended the quarter with $56.1 million of cash and cash equivalents. Deferred revenue was approximately $143 million and primarily reflecting customer prepayment associated with our NC-1 site and cloud services deployments.

During the quarter, we added approximately $83.2 million of project level equipment and bridge financing to support the continued development of our colocation and cloud services infrastructure. As Sam mentioned discussed earlier, completing the permanent financing for NT1 will further strengthen our financial capacity and allow us to recycle capital into future development. Overall, the quarter reflected continued positive adjusted EBITDA and substantial investment in infrastructure supporting our contracted growth. We remain focused on converting the investment into recurring revenue and cash flow while maintaining discipline around capital deployment. I will now turn the call back to Sam.

Thank you, Justin. Before we open the call for questions, I want to leave you with a few thoughts. Last quarter, we said the pieces of our development model were beginning to come together. They are. Since then, NC1 has moved into active customer deployment and toward full contracted operations. We've also focused our pipeline on the opportunities best positioned to move forward.

Importantly, we have recently entered into exclusivity with a consortium of well-known lenders for the proposed secured financing for NC1. The parties have commenced diligence, are negotiating definitive documentation and are working toward closing subject to customary approvals and conditions. This financing process has taken longer than we initially anticipated. But finally, reaching exclusivity and negotiating definitive documentation represent meaningful progress. If completed, the financing would return a significant portion of the capital invested in NC [Audio Gap] it would also allow us to advance the next site in our pipeline. But as my lawyers have advised me to say, there could be no assurance that the financing will be completed on favorable terms or at all.

This financing would also complete the first turn of the development flywheel we've described. We acquire Power Advantage infrastructure, we secure long-term customer commitments. We develop and stabilize the asset. We then access institutional capital and recycle our equity into the next project. Completing that first turn would represent an important inflection point for our colocation business.

We believe our next opportunity is also becoming increasingly tangible. Several sites have advanced significantly through our diligence process. Among the most actionable is a site that could support approximately 60 megawatts in 2027 and scale to more than 250 megawatts over time. The site has passed substantial diligence. We are now actively negotiating a purchase agreement as we complete the final stages of our evaluation.

Power available at this scale in 2027 is scarce. Our retrofit-first approach can bring capacity to market faster than traditional greenfield development, creating a meaningful speed-to-market advantage in a supply-constrained environment. We believe this combination of scarcity and speed to market should support premium economics. Across our pipeline, we're increasingly prioritizing opportunities with investment-grade credit support. We believe this will enhance project finance ability and execution certainty. This reflects the same disciplined sourcing approach that produce attractive economics at NC1, advantage power, speed to market and strong customer demand. We're not pursuing growth for its own sake. We're focused on opportunities that combine advantaged power, credible customer demand and financeable contract structure. Completing the NC1 financing would strengthen our ability to act on opportunities that meet those standards.

In cloud services, we're also converting strategy into signed contracts. The multiyear agreements we've signed since our last earnings call meaningfully expand our contracted revenue base and improve revenue visibility. These deployments are structured around firm customer commitments and are designed to be funded through customer prepayments and third-party equipment financing. This limits the capital required from white fiber while allowing us to retain attractive economics.

For the most part -- excuse me, for most of the past year, we've been building the individual pieces of this strategy. We're now beginning to demonstrate how they all work together. In colocation, we're moving towards a repeatable model for developing and financing long-term contracted infrastructure. In cloud services, we're pursuing longer duration customer engagements designed to generate attractive returns with limited white fiber capital. This is still -- there is still important execution to be done ahead of NC1 and on the financing, but completing this first turn of the flywheel would position us to enter 2027 with greater financial capacity, a larger contracted revenue base and a more actionable development pipeline. We remain focused on execution, capital discipline and building durable value for our shareholders.

With that, we're ready to take your questions. Joining us today for Q&A are White Fiber President, Billy Krassakopoulos; Chief Financial Officer, Justin Zhu; Eric Lang, an adviser to White Fiber and our former -- and of course, our former Chief Financial Officer; and Michael Francisco, Vice President of Cloud Services. Operator, please open the line.

[Operator Instructions] While we'll take our first question from Nick Giles with B. Riley Securities.

It sounds like demand is really strong for the remaining available capacity at NC10. Just hoping you could speak to that commercial process and kind of when you would ultimately cut it off or if you would be willing to kind of entertain other potential counterparties at this point?

Billy, would you like to take that? Sure. Thanks, Sam. Nick, we're still in the early phases of that. It's still a little early to comment on timing of when we would be able to set that up for any clients right now.

Fair enough, Bill.

It is -- I was going to add to that, but go ahead, Billy.

It is imminent. I mean we're fully focused on, like Sam said, completing that first turn of the flywheel and Phase 1 of North Carolina. But the next step is marketing and putting together a full project plan for Phase 2.

Yes. It's just worth mentioning and reiterating to Billy's point that these are -- we have wonderful champagne problems for Tranche 2. We have overwhelming demand for that. We do have a notification. We have our -- we have a legal obligation for Nscale to have to just notify them on the second tranche. But every counterparty is certainly looking at that. second tranche, and they've seen what we've been able to do already with the first tranche, and we've proven ourselves over and over again on how to get things done on time and within budget. So we're -- we'll be playing catch on the demand. We'll make sure that the economics are as premium as they can be for white fiber.

Great. I appreciate that. And then maybe just on the new site side, I was curious when you eventually acquire the site, kind of where it stands today, what type of development work you would be willing to complete before any commercial signing just to ensure that '27 delivery.

Bill, do you want to take that again?

Sure. We're looking at similar situations to North Carolina One, buildings that we can go into quickly, retrofit them. And I mean, our key advantage here is speed for ourselves and for our clients as well. The quickly we develop -- the more quickly we develop these properties, the more quickly we get clients in them, it serves both purposes. But the overall strategy that we're looking at is very similar to what we've accomplished at our North Carolina 1 facility.

We'll go to our next question from Greg Lewis with BTIG.

I was hoping we could talk a little bit about the cloud service business. Congratulations on bringing on a couple more customers. One of the things we've been hearing is that there's ample opportunities to bring on prepayments. How do you balance those upfront prepayments as you're thinking about your structures versus the overall return on, say, a multiyear cloud service business? Just trying to understand, I guess, what the hands worth going to bush, but just kind of curious how you're thinking about that as we continue to build out the cloud business.

Glad you asked that question. We have Michael Francisco, who is in the weeds of all that on the cloud side. Go ahead, Michael.

Thanks, Sam. So the way we think about this is we evaluate all of our deals at the project level and look for healthy terms across the life cycle of that deal. When we started kind of restructuring the cloud business earlier this year, we set some -- we set a framework around that, that really forced us to think about how we run this business in a way that might be a little bit different from the rest of the market and really focusing on high-quality customers as well as deals that have a positive cash flow throughout as well as deals that limit the amount of capital that we have to take out of our own funds in order to make those deals happen.

So when we think about the prepayments, that is a mechanism that we can leverage in order to help reduce the amount of capital that we deploy in support of these deals. And then we look at the structure of the deal across its life cycle to ensure that it is cash flow positive and get creative around the last couple of years on those deals. So you noticed as an example, Base 1 has a 2-year option for the customer to extend that deal. We can be creative with how we structure that to both benefit the customer over the term of that deal, but also allow us to adhere to the parameters we set out upfront.

Okay. Super helpful. And then my other question was around the pipeline. I guess the way -- like how do we think about with the growth pipeline, some of these projects where from a colocation standpoint, I'm assuming that those megawatts are just going to be bigger when -- versus, say, sites where we're going to use cloud services. But really, what I'm wondering is, could we see opportunities over the next couple of years where we're using a colocation customer, but also at the same location, maybe not in the same buildings, running GPU as a service. Is that something how we're thinking about maybe scaling that business also?

Yes, in fact. Michael, you should take this -- take that answer, but we'll be doing that, I think, sooner rather than later. Go ahead, Michael.

Yes. Getting to the vertically integrated model has always been the goal. And so to make that happen, we have to have our development pipeline on the data center side align with our customer pipeline on the cloud services side and have customers of the right quality that will allow us to get the right cost of capital to make that an interesting arrangement for us. We -- Billy and I have been talking about how we go get this done. And I think the time line on this has gotten shorter versus longer. And so ultimately, I think we will see a move to that. I don't think we will see all cloud services roll into the data center business, and I don't think that all customers at the data center level will come from the cloud business, but we'll be opportunistic about how we do that and look for other opportunities such as the arrangement with to identify and build capacity for our customers that we don't place in those data centers. It is something we want to get to, and we're getting there more quickly than we thought.

Super helpful. And congrats on getting NC1 off and running.

We'll next go to Raimo Lenschow with Barclays.

Congrats from me as well. That's amazing progress across the board, actually. I have 2 quick questions. One is, if you think, Sam, it's for you, as you think about the business, how do you think about the mix that we should think about in the long run between cloud services, colocation, -- you talked about managed services as well. on the last answer, it sounds like the first 2 are related, but how do you think about the evolution of the mix there and the pros and cons? And I had one follow-up.

Is the question, what do I -- how do we think of the colocation business and the cloud business and the pros and cons of mixing the 2 rather than keeping them separate? Is that the one?

And pursuing one more than the other, like just completely 2 slightly separate ways of doing the work.

Well, aside from -- I'd love Michael to add to this, but aside from the market assigns different multiples. to colocation versus cloud, and they both require different types of expertise and skill set. We have 2 very separate teams working on those businesses. Our colocation team is very separate from the cloud team and vice versa. And that's by design because, again, it's a very different skill set.

And integrating the 2, I sometimes wonder if that will affect multiples, but there's also a big reason to integrate the 2 because in a way, we can sort of double dip into the margin into the profits of the revenue. So maybe, Michael, if you want to add to that. But I know it's an ongoing debate that we have internally, and there is a path towards integrating the 2, which is what we're thinking about doing soon. I would love to hear from Michael if he has additional thoughts about that.

Yes. I think that there are benefits to both models there. But the way that we have it structured today, I actually think creates a good tension -- a healthy tension within the business. In order for the cloud business to be able to become a customer of the data center business and partner on those vertically integrated projects, we almost have to earn that opportunity. We are not -- the data center team is not beholden to the cloud organization to put our customers in those locations.

Instead, we need to have a compelling customer and a compelling economic case in order to displace some of the demand that they already have. And I think that tension is very healthy because it allows the cloud team to have some goals and some parameters around how they're going to actually be able to do that vertical integration. So again, I could see benefits to both sides, but I think the healthy tension inside the organization and the desire for the cloud team to partner more closely with the data center team on specific projects is good for all of us.

Okay. Perfect. And then can you talk to the managed service approach? Like how -- I mean, obviously, it helps you a lot on not having to deploy capital, et cetera. But like what sort of a margin profile that you -- that we should think about there?

The managed services model is pretty interesting. It was actually something that we put together in anticipation of internal enterprise adoption of AI for R&D and kind of internal development projects. What we found as we put that together is that the demand for it is much greater than enterprise customers who traditionally want to actually own the CapEx expense for a number of different reasons. We're actually seeing a lot of demand both from other Neo clouds, interesting financial partners as well as some of the frontier and AI labs that are -- or have grown to a point where they're actually considering taking on some of the CapEx.

From a margin perspective, we see a very healthy margin on those. And I'm going to pause for now on talking about margin until we have some other things to talk about around this. But one of the key benefits, along with the capital-light kind of deployment opportunity is that we start driving a meaningful margin from day 1. So because we are not working on having to pay off finance debt or data center costs, all of the revenue that we produce as part of those deals starts evenly creating revenue for the company from day 1 throughout the life cycle of the deal.

And then on top of just standard development and operations, we have adjusted the way we think about software development, focusing our internal software development primarily on projects that help us drive revenue and balance sheet growth. So we have a road map focused on how do we drive better efficiency to reduce costs, and we have a road map to think about how we drive incremental value for customers and revenue inside of those deals and deployments.

On top of that then, we layer on third-party services to provide an appropriate toolbox of APIs for developers that are leveraging our bare metal solutions. So it allows us to have layers of sales opportunities on top of each of these managed services deals that create incremental revenue and incremental margin and significant value for our customers. In general, the margin profile would look more like a software offering than a hardware offering.

And we'll take our next question from Brian Dobson with Clear Street.

So you've been signing a significant number of contracts and agreements over the past quarter. I guess as you're in discussions with those clients, do you find that they're leaning more toward longer duration contracts? And I guess, how is the execution that you've done at NC1 impacting those conversations? It must be a net benefit?

I don't think the execution at NC1 is related to the cloud contracts that we've been signing up. They're 2 separate businesses. But maybe, Michael, do you want to just speak about the longer duration contracts on the cloud side? And I'm happy to double-click on if there's anything unanswered, Brian, feel free to ask away. I just want to make sure your questions are answered. Go ahead, Michael.

On the cloud side of the business, we're seeing an interesting dynamic with regard to term of contract. Last year and kind of coming into this year, customers were often looking for shorter-term arrangements. But with the dynamics of the pricing that we're seeing inside the cloud GPU model, I think customers are starting to reevaluate how they procure these and on what duration. If we look at H100s are probably the best example, given they've been in the market the longest, the actual cost per GPU hour for those is actually higher today than they were when they released into the market. And so customers are looking at some of those dynamics and then considering what is their total cost of ownership or total lease cost across the life cycle of those deals and looking to both, a, preserve their access to those GPUs, so they're not having to go out and fight a tough market 3 years from now to go find new capacity, but also looking for ways to bring that cost down and ensure that they have access to those GPUs for as long as makes sense for them.

And so those longer duration contracts also allow us to adjust the economics within those deals in a way that's favorable to the customer and actually helps improve our margin throughout the life cycle. So there's a couple of dynamics there, both in terms of in the market and kind of how customers are thinking about these things that are influencing these longer-term deals.

Yes. Great. And then you mentioned that your next opportunity is now in late-stage diligence, I guess. Can you give us a little bit more color on what that might look like? And at this point, given your experience, would you favor like a single campus opportunity or a multi-campus portfolio type of development?

Yes, those are great questions. Billy, do you want to take that?

Sure. So it's leaned towards a single tenant opportunity. And again, think of the process that we went through for North Carolina One, kind of the same deal here, final stages of due diligence. We'll be marrying the opportunity to a client shortly if everything goes well with the technical due diligence and executing the same game plan that we've done and are completing right now in North Carolina One.

We'll take our next question from George Sutton with Craig-Hallum.

So a lot of great updates. I wanted to take a higher-level view on one of the challenges in the market recently have been the not in my backyard theme. And it would seem to me that retrofits along with your cross DC initiatives would be great answers to that challenge. Can you just talk through that in terms of the things you're looking at?

Yes, certainly, and Billy will talk to that just before he starts, I do want to embrace how important it is to engage with the community. We had a Community Day at North Carolina One, and we discussed how we're taking 85% less water than the previous tenant that there'll be less noise than the previous tenant. The retrofit format really does solve a lot of the pushback related to data center build-out, particularly against Greenfields. And so we think engaging with the community, getting support from local community leaders using retrofit as opposed to greenfield really does help solve a lot of the problems that are being discussed right now nationally. In fact, CNN just published a pretty long interview about WhiteFiber's approach and how it's uniquely designed to help mitigate all the pushback. I would encourage all the listeners to listen in on that. Billy, do you want to add a few more points on that question?

Yes, for sure. I mean, like Michael said, the GPU division and the data center division work hand-in-hand and the cross data center platform is going to enable us to deploy that technology in our sites and all the projects, all the stuff in our pipeline that we've been presenting have always been deemed on the smaller side, 30 megawatts, 60 megawatts, 99 megawatts. There's a reason for that. It's done on purpose. Like we're able to execute these sites much quicker and bring that capacity online much quicker than competitors can, like we've proved with North Carolina One. And with the cross data center platform, we can bundle these sites into larger clusters. So all this stuff is work in progress, stuff that we're cooperating together with the GPU division and the data center division, and it's all strategic.

Yes. I want to highlight that point that Billy made. This technology that we patented was patent pending is very unique. And the reason for it -- the impetus for it is if we have these smaller modular sites and if we can create a super virtual super cluster on these modular sites, we can basically solve for disparate smaller sites and just create them into these virtual larger ones under one logical system. And so this technology helps basically solve for that issue. And if we have smaller sites, we could just connect them through this technology, which we think is transformational for the industry, let alone white fiber.

Makes great sense. One other question on Krambu. So if I understand it, you have worked out access to 100 megawatts in 2027 with Power Access. Can you just walk through how that's influencing what's in your pipeline and how you're -- just a little bit more on that deal?

Yes. So the genesis for this, we have been working with Krambu for quite a while, and we have some -- a couple of deals in late stage that we hope to be able to talk about where we're leveraging their facilities. Both Krambu and ourselves see the market today is trying to solve GPU and power access as 2 separate problems. And it's really hard to get those time lines to match with customer demand and their own time lines and goals and priorities around deploying the infrastructure.

By partnering together, we create a single phase, and we can align our -- both of our sales pipelines as well as our supply chain pipelines together in order to provide very clear views into what's available and when for our customers. The genesis of this actually started as a technical collaboration because as we've shifted our focus to the physical layer of GPU infrastructure, we have been working with them around how do we think about driving the most value out of these highly dense GPU clusters. And so for us and them being able to bring the skill sets of our engineering teams together as well as align on the availability time lines.

We have a very compelling kind of offering and workflow for customers who are looking to do planning going forward, particularly those customers who have, I would say, defined scaling plans where we can align their needs with what we have upcoming. From how it has affected our pipeline, it's actually what it has done today because we're just announcing this. So we'll see what it does relative to growing the pipeline. But what it does is it makes some of our pipeline more realistic in terms of being able to fulfill those customer demands, both more near term, but also long term. We had the good fortune of expanding a couple of customers this quarter, and we have others that we're talking to about working with them to partner across the next 12 months around their deployments and what they need and an arrangement like Krambu allows us to give them a very clear time line and be able to fulfill and act as a good partner for them to help fulfill those needs.

We'll take our next question from Paul Golding with Macie Capital.

Congrats on all the progress. Just wanted to ask initially on the 200-megawatt incremental opportunity at MC1 that Duke is evaluating. How would you expect to see that capacity potentially come on? Would it be phased once again as with the first around 100 gross? Or would you expect to see this load study potentially lead to the full 200 coming on all at once? And then how might you market that? And then I have a follow-up.

Billy, you should definitely take that one.

Paul, a little difficult to say right now. It's still too soon in the game for that. There will more than likely be one or even multiple sites that we're enacting on before that 200 megawatts is approved or any schedule for deployment on that is to be released.

Understood. And then maybe from a cloud perspective, just turning to the comments on the prior question. I just wanted to double-click on the GPU availability itself. We've been seeing in the marketplace, of course, how constrained GPU capacity is through this partnership that you have, just in terms of access to the compute, how -- I guess, how confident are you into sort of the forward-looking availability of that supply, in particular, as you noted that you have a new cloud services agreement around Vura Rubin and V300 infrastructure, maybe in particular around Vura Rubin, given the memory considerations there and presumably the opportunity to sell through at a higher price to your customers. Just how is the absolute quantum of availability of compute looking based on the relationships that you have?

So I'll address this in 2 parts. So the first part is relative to our relationship with Krambu, it actually helps us with availability, being able to leverage both our connections kind of across the industry from OEMs and NVIDIA as well as their own allows us a little bit better access. And I think some of the work they've done around cluster density and kind of the footprint and how their data centers are designed have helped them obtain some allocations that we can then take advantage of. So on the whole, it is a net benefit to us to be able to have this partnership relative to allocations.

A second piece, I think I would like to address on this is that one of the things that we did when we went to restructure this business is to put a strategy in place that prevents us from chasing our tails around near-term GPU demand. It's funny, one of the things that our customers love about us is that we're very willing to say no. And the reason we say no is because we do not want to be able -- we do not want to put ourselves or our customers in a position where we are making commitments around delivery time lines that we are not absolutely certain that we can meet. And so a lot of the discussions we're having are for deployments that are far enough out that we can be very certain that we can get the GPU allocations in place in order to serve those customers. And particularly for those customers that are partnering with us to scale, we are looking out at a time line of about 12 months. And so it's a little bit easier for us to make sure that we can source the right infrastructure in order to support those customers because of the way that we've set the business up.

Maybe just to sneak one in on the back of that response. Do you -- might we expect to see White Fiber buying GPU compute speculatively in the marketplace based on your visibility to this sort of demand curve and how it's coming to you and realizing committed contracts?

Perhaps, but I don't see that happening in the near term. It would have to be really the right opportunity for us. If I go back to the initial parameters that we put in place as we started restructuring the business, that speculative purchasing doesn't really fall into that. Now there are some things that we're working on where it might make sense for us to do that in the future, but I'm not ready to comment on kind of what those developments are. But I will say that today, we are really focused on fulfilling real customer need based on real contracts. with high-quality customers. And so the speculative purchasing is not a thing that's on the table for us at the moment. Priority given our available capital today.

We'll take our next question from John Todaro with Needham & Company.

I was hoping to just get a little bit more commentary on the financing market, whether for NC1 or some of the future sites, the most likely guarantors here. We talking about chip manufacturers? Are the banks starting to step in more? In the future, do you target more hyperscaler leases? Just any commentary there.

I mean, overall, I mean, we're really just trying to solve for the lowest cost of capital. And I think we've had certain things that we've learned along the way in this NC1 financing process that, as noted, has taken longer than expected, and there would have been certain contractual features certainly that would have made it easier to finance. So we know what we need to solve. We know that we want to establish a structure with counterparty that's financeable from day 1 and have that firmly underwritten. So whether that be directly with a hyperscaler, backed by a chip manufacturer, whatever structure is really advantageous from a cost of capital perspective. That's kind of our priority. So there's not one above the other. It's really just solving for the lowest cost of capital and the best financing structure.

Okay. Understood. Makes sense. And then one on the Cloud segment. It seems like pretty good pricing on all of them. The GPU per hour pricing I was backing into was ranging from $3.47 up to above $7. And I think that lower one includes a revenue share, so it likely comes even better. Just wondering if margin starts to expand maybe sooner than folks were initially thinking, I would love to just get your thoughts there on that margin expansion opportunity.

So from a margin expansion standpoint, as we look across the deal types, I think we will see an improvement in margin kind of across our deals as we look at the mix that we're looking to deploy over the coming years. The customers that we talk to are especially these folks that are more mature in the AI lab space or in other areas or particularly in the enterprise, they understand the balance between cost and quality and overall ROI on those clusters. And so many folks have been burned kind of going for the cheaper pricing and winning -- in doing so, they actually get less value out of that because they suffer from more downtime missed SLAs and those sorts of things. So because of the quality of our engineering team, we are able to deliver really high-quality deployments. And as a result of that, that's the thing that customers are willing to pay a bit more for. So I think we will see improvement for a couple of reasons over time, but it's going to take some time for that to develop.

We'll next go to Michael Donovan with Compass Point.

Congrats on the progress, guys. So you now have operations across the U.S., Canada, Iceland and France. How do you think about geographic expansion from here? What other markets look most attractive? And what factors are driving where you may potentially expand?

Those are probably different answers depending on the business unit. Billy, do you want to talk about it from a colocation perspective and then Michael can discuss it.

Sure. On the data center colocation side, it's quite simple. It's a mix of client opportunity and where we can marry that and match that with available power in the time line required. So everything that we're looking at right now is mostly in the United States, a little bit in Canada for the data center colocation side.

On the cloud side, it's -- pardon me, sorry, you there. On the cloud side, it's really driven by customer demand. And customers have different requirements for where they want their GPUs to be placed, whether it be for compliance reasons or other things, performance due to low latency, things like that. So it's really driven by our customers. And so most of our deals that we're seeing today are in North America, but we are seeing some interest in European deployments, particularly for folks that have GDPR concerns, things like that. So we'll see how that goes over time. But right now, the economics of where we place those GPUs also has an impact. And right now, the U.S. is the most attractive market relative to what we're seeing from our customers.

That's helpful. And I understand you want to keep discussions at a high level, but on Project Redwood and cross data networking, is this primarily suited for inference? Or could it support training as well? And theoretically, how many geographically separate sites could you aggregate?

Today, the way that we're building it is to support multiple use cases. One of the reasons why we selected Modal as the customer for our R&D cluster is because they were doing some intense training workloads. And those workloads being able to put those clusters and run them as a single virtual cluster, training is the most demanding workload that we could put on there. So I think we can support kind of both ends of the spectrum from training to inference, and it will be interesting to see what kind of use cases our customers want to use that for because while we talk about it as a single virtual cluster is kind of the primary way people are talking about that today. There's lots of other potential use cases we can look at from telecommunications to edge computing and others. So we'll be really interested to see what our customers want to do with that once we bring that to market and finalize the testing across the full fiber span.

And our last question comes from Nathan Francovitz with Cantor Fitzgerald.

Just on the cross data center product, I guess, how do you think about the long-term opportunity? And what's the pathway to monetization? Is it more of an internal capability? Or can you talk through like the potential scalability for commercialization?

Could be both internal and we're even thinking about licensing it, but Michael is leading that work stream. So go ahead, Michael.

Thanks, Sam. Yes, Sam is correct. I think there's -- and as Billy mentioned earlier as well, we think there's some really compelling use cases for our own internal utilization of this, whether it be connecting 2 large sites with lots of megawatts. There's clearly some economic value in being able to connect as an example, to 50-megawatt sites and go market that as 100. There's some value in that. There's also value in being able to leverage fragmented power resources.

That's an understatement. There is massive value in that.

So I think as power continues to become a challenge for folks, being able to aggregate multiple sites into a single cluster is going to be a value proposition. And we don't know yet what's going to happen from a legislative standpoint. But say we get to a world where there are taxes applied to clusters -- or excuse me, data centers over certain megawatts, this technology allows us to really think strategically around how we deploy it in order to preserve kind of the economic value of the sites by kind of bringing them together. In terms of the number of sites, we will see. The next phase of testing around this will be to test and demonstrate the efficacy of the hub-and-spoke model. And so we believe that it can expand significantly, but we will wait to see when we have some real data that we can present to the market.

I'd now like to turn it back to our speakers for any final or closing remarks.

Sorry, I was on mute. Thank you, everyone, for joining us today. We look forward to the next quarterly call. Until then, we'll be working very hard and delivering results. Thank you.

Thank you. And this does conclude today's call. We thank you for your participation. You may now disconnect.

Hello, and welcome to the WhiteFiber First Quarter 2026 Earnings Conference Call. Good morning, and thank you for joining us. We will begin with prepared remarks from management. [Operator Instructions]

As a reminder, today's conference is being recorded. I'd now like to turn the call over to your host, Cameron Schnier, Vice President of Capital Markets and Corporate Strategy at WhiteFiber. Cameron, please go ahead.

Thank you, and welcome to the WhiteFiber First Quarter 2026 Earnings Call. Joining me today are Sam Tabar, our Chief Executive Officer, and Erke Huang, our Chief Financial Officer.

Before we begin, I'd like to remind everyone that some of the statements we make on this call are forward-looking in nature and subject to risks and uncertainties that could cause actual results to differ materially. Such risks and uncertainties include, but are not limited to, those factors described in today's earnings press release, our Form 10-Q for the quarter ended March 31, 2026, filed today as well as our other filings we make with the SEC from time to time. Our remarks today may also include non-GAAP financial measures. Reconciliations to the most directly comparable GAAP measures can be found in our Form 10-Q and in the earnings press release posted on our website.

Following our prepared remarks, we will open the call for questions.

With that, I'll turn the call over to Sam to discuss our performance. Sam?

Thank you, Cam, and thank you, everyone, for joining us. The first quarter was another important quarter for WhiteFiber. We delivered a solid first quarter with year-over-year revenue growth, strong gross margins and positive adjusted EBITDA while continuing to invest in the AI infrastructure platform we are building.

More importantly, we continue to make progress across the four areas that matter most in creating long-term value creation, data center development, customer demand, financing and cloud capacity deployment. The market backdrop remains very strong. Demand for AI infrastructure continues to exceed the available supply. Customers need power. They need high-density capacity. They need speed and they need partners who can actually execute.

That last point is important. In this market, demand is not the main constraint. Access to potential site is not the main constraint. The real constraint is execution. Power has to be secured, equipment has to arrive, capital has to be available. Customer requirements have to be finalized. Construction has to be managed and the facility has to be delivered and operated at a high standard. That is where we believe WhiteFiber is differentiated.

Our retrofit first strategy is designed to reduce the development risk and shorten the path from site acquisition to revenue. Montreal-3 is a good example. We converted an existing industrial facility into a Tier 3 data center in approximately 6 months. Who else does that? And now the site is supporting Cerebras.

At the same time, the current market environment is challenging. Equipment lead times are longer. Supply chains are tighter. Utility and commissioning time lines are complex. That is why we are disciplined about how we commit to new projects and why execution remains the central focus of the business. With that context, I'll start with NC-1, our site in North Carolina.

NC-1 continued to make substantial progress during the quarter and after quarter-end. Duke Energy has completed delivery of the initial 54 gross megawatts of utility power to the site, supporting the first 40-megawatts of IT load under our agreement with Nscale.

Construction and commissioning activity remains highly active with approximately 600 personnel on site last week as the facility moves through final commissioning stages. The major equipment packages needed for the deployment, including generators, UPS systems and chillers are all on site. The remaining supply chain item we are managing relates to certain medium voltage switchgear components. That is not a broad equipment availability issue across the project. We expect to begin delivering initial capacity within the next couple of weeks.

In connection with the supply chain-related timing item, similar to what we are seeing across the broader sector, delivery may begin slightly later than June 1. Based on our current discussions, we do not expect a material delay, a material impact to our customers' commissioning process or a change to the overall project economics. The core commercial rationale for NC-1 remains unchanged. The project is backed by a long-term agreement with Nscale and the deployment is supported by Nscale investment-grade hyperscaler offtake.

As a reminder, construction started in January. We are moving this project from construction to initial capacity delivery on a time line unmatched by any of our peers for a 40-megawatt AI infrastructure deployment. We believe NC-1 validates the strength and speed to market of our retrofit model, especially in an environment where utility time lines and electrical equipment constraints are impacting projects across the broader data center industry.

Importantly, NC-1 remains a strategic platform asset beyond the initial deployment. We continue advancing plans for additional capacity at the site and expect to begin marketing the next 45 megawatts tranche of capacity this summer. We are also working with the utility on longer-term power expansion opportunities with the potential to scale the site up to approximately 300 gross megawatts over time. That is why we remain excited about NC-1.

The initial Nscale scale deployment is important, but it is not the full opportunity. We believe NC-1 can become a larger platform asset over time as we bring the first 40-megawatts online, advance the next tranche of capacity and continue working on the broader power expansion path. In short, NC-1 is nearing the point where we could begin converting contracted demand into revenue. We believe this will be an important milestone for -- we believe that this will be an important milestone for WhiteFiber and a major proof point that our retrofit model can create large-scale, high-value AI infrastructure.

Turning to Montreal-3. The first quarter was the first full quarter of operations for that facility. Montreal-3 is supporting our colocation agreement with Cerebras. Cerebras is an important innovator in AI infrastructure, and we are proud to support their growth. We congratulate their team on this exciting milestone as they become a public company, and we look forward to continuing to support them as their infrastructure needs grow over time.

After quarter end, we completed the purchase of Montreal-3 through the exercise of our previously disclosed purchase option. The purchase was supported by our amended credit facility from the Royal Bank of Canada. We believe owning Montreal-3 is strategically beneficial. It reduces our lease payments by approximately CAD 3.1 million or USD 2.3 million annually over the remaining term. It also gives us greater control over a strategic asset that is already generating revenue, and we believe it allows us to capture more of the upside if the site can be expanded.

On that note, we have submitted an application with the local utility to more than triple the available power of that site over time. This remains subject to utility review and approval, but it is an important part of why we believe ownership of that specific site is valuable. If approved, the incremental power would increase the strategic value of Montreal-3 well beyond the current deployment. This is a good example of how we think about our platform. We are not only bringing sites online quickly. We are also looking for ways to own, optimize and expand assets once they are operational.

Turning to Montreal-2. We continue to advance discussions around the best customer solution for that site. We have quality customer interest in the facility, and we believe the location can support valuable enterprise and AI infrastructure use cases. The site is smaller than in NC-1, but it has strategic value because of its location, connectivity and potential fit for customers seeking more targeted deployments. We are focused on matching the site with the right customer and the right commercial structure. As with the rest of our pipeline, we will remain disciplined and move forward only when there is customer demand, economics and capital plans are all aligned. More broadly, our pipeline remains active and continues to improve in both quality and scale.

We are not constrained by customer demand, and we are not constrained by access to potential sites. We have both. The gating item is making sure each project has the right combination of power, customer alignment, capital availability, equipment visibility and execution certainty before we commit. That discipline is important, but it should not be mistaken for a lack of opportunity. We are seeing more customer interest than we can currently serve, and we are evaluating sites that are larger and more scalable than our initial Montreal deployments. The opportunities we are most advanced on today are not small follow-on projects. Several are comparable to or even larger than NC-1 and potential scale.

Importantly, we believe these opportunities also offer meaningful near-term power availability, along with significant expansion paths over time. We are not disclosing specific locations or counterparties before transactions are finalized because that protects our negotiating position. But we do want shareholders to understand that the quality and scale of the pipeline has continued to improve.

Our goal is to demonstrate in the coming months that this pipeline can translate into actionable customer-backed projects. We are focused on opportunities where we can align customer demand, supply control, power availability and financing from the outset. That is important in this market. Equipment lead times are increasing, utility time lines are complex and customers are demanding more certainty before committing to large deployments.

We do not want to announce capacity, just to announce capacity. We want to build a pipeline that could be contracted, financed, delivered and operated reliably. We are advancing several opportunities and expect to close at least one additional site in the coming months, subject to final diligence, documentation, customer alignment and capital availability. That is the standard we are holding ourselves to. We believe this approach gives us the best chance to turn a strong pipeline into durable financed revenue-generating assets.

Turning to our Cloud business. We discussed last quarter, we made the decision to strategically pivot this business in Q1. This shift positions us to deliver a longer duration enterprise deployments, managed infrastructure services and next-generation GPU capacity. While we implement this strategy, it has created near-term revenue pressure, and we continue to expect the second quarter to be the low point for Cloud revenue.

However, we are seeing positive outcomes faster than we expected as a result of this change. The Cloud business today is in a much stronger strategic position than it was only a few months ago. We remain focused on improving our customer mix, extending contract duration, sharpening our return thresholds and moving away from shorter-term commodity bare-metal leasing. At the same time, we are seeing accelerating momentum in high-quality pipeline growth and deal velocity.

Notably, we are in the final stages of a long-duration 9-figure cloud opportunity with a high-quality enterprise customer in a new market. We expect to provide more detail if and when the agreement is executed and customer disclosure approvals are in place.

We view opportunities like this as important validation of our revised cloud strategy. They combine long-duration customer contracts, next-generation GPU infrastructure, customer supporting funding and attractive project level financing. They also diversify our Cloud footprint geographically and reinforce that demand for high-performance AI infrastructure is global.

In addition, we signed a 2-year agreement with Hyperbolic for approximately $17 million of total contract value supporting Modal Labs as the end customer. This deployment uses H200 GPUs from our existing owned fleet as part of our cross data center R&D project, so it does not require incremental GPU CapEx. The deployment is expected to begin contributing revenue in the coming months.

This was a competitive process with multiple customers interested in acting as a design partner for the ongoing R&D. We selected a customer who saw value in the novel infrastructure and how they can apply it to their environments. As a design partner, Modal Labs will support ongoing R&D through input on design, development and testing. We expect that they will scale with this cluster as they move through later phases of R&D. This structure allows us to monetize existing capacity while continuing to develop technology that we believe can differentiate our platform and contribute meaningfully to our revenue growth. We plan to announce the outcomes of this first phase of R&D in late Q2.

More broadly, we are seeing strong interest in current and next-generation GPU capacity. Our current prospects and customers are increasingly looking for reserved, high-performance infrastructure as well as vendors that can reliably support their scaling requirements. That is where we believe WhiteFiber can compete more effectively.

Our Cloud pipeline has expanded meaningfully. Today, we are tracking more than 50,000 GPUs, representing a weighted pipeline value of approximately $3.3 billion based on pipeline stage and probability of close. Not all of that will convert, and we will remain selective, but the size and quality of the pipeline as well as diversity in deal types shows the level of demand we are seeing for high-performance AI infrastructure. Based on our current momentum, we expect Cloud revenue to grow sequentially in the third quarter and build through the second half of the year.

We are focused on securing contracts that are meaningfully cash flow positive across the term. This allows us the fund growth through customer prepayments and attractive equipment financing rather than relying solely on the corporate balance sheet. This is important for capital discipline. It allows us to pursue Cloud opportunities where customer demand, contract structure and financing are aligned from the beginning.

We are also continuing to invest in technology that can differentiate our platform over time and is directly aligned to our go-to-market strategy. This includes our cross data center R&D work, high-performance networking, storage architecture and other monetizable innovations. We are making progress towards protecting this work through intellectual property filings, and we expect to share more detail through technical updates, announcements and white papers when appropriate.

This is the strategic role we want Cloud to play within WhiteFiber, a disciplined enterprise-focused platform that complements our data center business, deepens customer relationships and creates additional ways to monetize our infrastructure and technical capabilities. The key point here is that we are not chasing Cloud revenue at any price. We are focused on deployments that meet our return thresholds, including durable customer commitments, offer built-in scaling and strengthen the overall platform. We believe the Cloud business is now moving from a transition period into a renewed growth phase.

I will now turn over the call to our Chief Financial Officer, Erke, to discuss our financial results.

Thank you, Sam. I will now review our first quarter financial results and balance sheet. First quarter revenue was $21.9 million. This compares to a $16.8 million in the first quarter of 2025. This was an increase of 31%. Cloud revenue was $16.8 million. This compares to $14.8 million in the first quarter of 2025. Cloud revenue was lower than the fourth quarter as expected, as we continue to reposition capacity towards longer duration enterprise deployments. Colocation Services revenue was $4.8 million, this compares to $1.6 million in the first quarter of 2025. The increase was mainly due to Montreal-3, which began billing in October 2025 under our Colocation agreement with Cerebras.

Gross profit, excluding depreciation and amortization, was $13.2 million. Gross margin was 60.2%. This compares to a gross profit of $10.1 million and gross margin of 60.5% in the first quarter of last year. Depreciation and amortization expense was $6.4 million. This compares to a $3.8 million in the first quarter of 2025. The increase was mainly due to the expansion of our Cloud and Colocation infrastructure.

General and administrative expense was $17.8 million. This compares to a $4.2 million in the first quarter of last year. The increase was mainly due to the share-based compensation, higher headcount, costs related to operating as a stand-alone public company and continued investment in the platform. Based on our current expectations, we do believe Q2 G&A will decrease by around 20% compared to the first quarter.

Operating loss was $11 million. This compared to an operating income of $2 million in the first quarter of 2025. Interest expense was $2 million, primarily related to our convertible notes issued during the first quarter. Net loss was $12 million. This compared to a net income of $1.4 million in the first quarter of 2025.

Adjusted EBITDA was $3 million. This compares to adjusted EBITDA of $6 million in the first quarter of 2025. The year-over-year change was mainly due to higher operating expenses as we invested in the business and built the public company platform.

Turning to the balance sheet. We ended the quarter with $75.8 million of cash and cash equivalents and $4.3 million of restricted cash. The total cash and restricted cash was $80.1 million. During the quarter, we completed a $230 million private placement of 4.5% convertible senior notes due 2031. In connection with the notes, we also entered into a zero-strike cost structure. The structure is designed to materially reduce potential dilution from the notes. Cash declined from year-end primarily due to the continued capital investments in data center infrastructure and equipment deposits, partially offset by proceeds from the convertible notes financing and proceeds from the sale of certain GPU assets.

In March, WhiteFiber Iceland entered into a secured term loan facility with Landsbankinn. The facility provides up to $20 million of available borrowings as secured by WhiteFiber Iceland shares and certain assets, including GPU service and related equipment. After quarter end, we drew $18 million under this facility.

Also, after quarter end, we entered into amended credit agreement with RBC. This provides a CAD 28 million facility to support the purchase of the Montreal-3 site. The purchase closed in May, and we also continue to evaluate additional financing options to support our data center development pipeline. Overall, the first quarter showed continued revenue growth, strong gross margin and positive adjusted EBITDA. We also continue to invest in the infrastructure and balance sheet needed to support the next phase of the growth.

I will now turn the call back to Sam.

Thanks, Erke. Before we move to Q&A, I want to take a step back and talk about where we are as a company. The first part of this year has been about getting the platform ready for the next stage. At NC-1, we have been focused on bringing the project through construction, energization, commissioning and initial revenue. At the same time, we are well underway in a formal project-level financing process for NC-1. Lender diligence is active, including site level diligence, and we are very encouraged by the quality of engagements to date. We believe the process is moving toward a near-term financing solution that reflects the quality of the asset and the long-term contracted cash flow profile of the project.

This financing is important because it can validate the credit quality of NC-1, recycle capital already invested in the project and give us more flexibility to move on to the next customer-backed opportunity. We also continue to maintain access to bridge financing solutions. We view that capital as a flexibility tool, not the long-term financing plan.

Importantly, it can allow us to activate select near-term opportunities without waiting for permanent project level financing to be fully in place. That matters because some of the best opportunities require speed. So if we have the right site, the right customer and the right return profile, we want the ability to move quickly. The goal remains the same, complete NC-1, put efficient project-level financing in place and recycle capital into the next customer-backed opportunity. We believe that model is now coming into focus, and we are closer than ever to demonstrating the flywheel we have been building toward. That flywheel is straightforward, secure a strategic site, match it with a high-quality customer demand, finance the project efficiently, deliver the capacity and recycle capital into the next opportunity.

We know shareholders want more detail on the pipeline. We want to be transparent, but we also need to protect the company's negotiating position. If we identify specific sites, markets, sellers or customers before transactions are finalized, we can give counterparty leverage and make those opportunities harder or more expensive to close. That would not serve shareholders. What I can say is that customer demand is very real, and it is not concentrated in one conversation. We are in advanced discussions with two large high-quality customers for two distinct site opportunities, including customers with scaled AI infrastructure needs and strong credit profiles.

We also continue to see broad demand from a wide range of credible potential customers across the market. The Cloud business is also in a much better position than it was only a few months ago. The reset is working. We are seeing better customers, longer commitments, stronger financing structures and more opportunities supported by customer prepayments and equipment financing.

So while the first part of the year required a lot of foundational work, the pieces are now coming together. NC-1 is nearing initial revenue. The financing process is advancing. The pipeline is larger and more actionable. Montreal-3 is operational and now owned. And Cloud is moving from transition into renewed growth. Our job now is quite simply execution. That means delivering for customers when projects are complex. It means solving problems, staying aligned and giving customers confidence that WhiteFiber can support their growth. That is how one project can lead to the next. That is what we are building at WhiteFiber, a project that can secure strategic sites deliver high-quality infrastructure, support customer growth, finance projects efficiently and repeat that model over time.

As a note, WhiteFiber President, Billy Krassakopoulos; and Head of Revenue, Billy Cladek, will be available for Q&A today. Yes, that is two Billy. Billy Cladek joined WhiteFiber earlier this year and leads enterprise and partnership strategy for our cloud business. He brings more than 12 years of experience building and scaling go-to-market organizations across SaaS, commerce and AI infrastructure. More recently, he led global go-to-market at Firework, where he helped scale the company's commerce platform across major global retailers and consumer brands.

Thank you. I will now open the line for questions.

[Operator Instructions] And we'll go right to Nick Giles with B. Riley Securities.

Really, just my first question was about the expansion at MTL 3. I was wondering if you could talk about the CapEx side. Would we expect the expansion to benefit from the same level of CapEx savings? And then just any sense on timing as far as utility approvals or when that kind of additional build-out could ultimately begin?

You got it. We have two Billy on the line. So I'm going to refer to either Billy, Data Center or Billy Cloud. So Billy Data Center, you want to go?

Nick. Yes, the real key gating item here is power. We've made -- since we've acquired the building, we made the application to the utility provider for a threefold increase. And all indications are positive at the moment, but we don't have any news around timing right now. But if and when we receive additional power, Cerebras will be certainly one of the first customers we speak with. They're an important customer for us, and we welcome the opportunity to support more of their growth at Montreal-3. And to your question around CapEx, again, it's a little early to pronounce any sort of budgets or stuff like that, but we do expect them to be in line with our past Montreal build.

Understood. That's very helpful, Billy. And then just maybe on the point with Cerebras. I mean that agreement is in line with your thesis around the importance of inference. So as we think about new site selection, can you just speak to how you're balancing sites that could be closer to large metros that may be better suited for inference or if you're looking more to kind of Tier 2, Tier 3 markets and how ultimately acquisition costs would differ in either scenario?

For sure, acquisition costs are a little lower in Level 2 type markets, but everything we're looking at right now is to support inference-type architecture and solutions. And it's really active conversations that we have with customers in our pipeline.

We go next to John Todaro with Needham & Company.

Congrats on all the progress here. Certainly, a number of items going on. So I guess, first, on the pipeline, when you said expect to close one new site in the coming months, assuming that's kind of land in power, I know in the past, you've kind of -- as you thought about acquisitions for additional sites, there's been kind of some negotiation with a potential counterparty on a lease kind of in the background. So when you do announce the new site, should we be thinking, hey, a lease comes in short order after that or not? Would it actually take some time to then go out and market that site?

I would like Cam and Billy to answer that question.

For sure, a major part of our site selection process comes from the dialogues we're having with current customers in our pipeline. So yes, we do -- we are targeting sites with more or less a customer already -- customer engagements already happening or a customer already in hand.

Yes. I mean it would probably somewhat depend on the materiality threshold and when we actually announce the acquisition, whether there's any delay, but that's certainly our goal, our intent is to pair them as soon as possible.

That's very helpful. And then maybe just a sort of housekeeping item. That 9-figure Cloud contract that you mentioned, I'm sorry if I missed it, where do we expect that to be slated? Is that something part of the pipeline or somewhere in an existing site or a third-party data center provider?

We can't say too much about that until that thing is signed. So unless Billy Cloud wants to give some more color on it. But my -- we would like to just keep our -- we just want that thing signed before we give out any information, if that's okay.

Generally, you assume contracted. We didn't announce that it would be in one of our own sites, so you can directionally assume it would be a third-party DC.

Our next question comes from the line of Paul Golding with Macquarie.

Congrats on all the progress. I wanted to ask about GPUs initially. Noticed in your press release that the Hyperbolic to your agreement was for existing H200s. I was wondering if the conversations you're having incrementally beyond the Hyperbolic deal with potential counterparties around Cloud is also relating to that generation of GPU? Or if we may expect that you might have to acquire newer iterations of accelerators as your conversations progress around Cloud? And then I have a follow-up on data centers.

Billy the Cloud, do you want to take that?

Yes, happy to answer that. Given the dynamic nature of this market, we look to achieve a few things with all of our deals, and that's a minimum 3-year term cash flow positive for the life of the deal and prepayments that prevent investing capital from our balance sheet. With that said, most of the deals that we're looking at right now are for generations beyond the H200, mostly the B300s that we're seeing out in market.

And as we continue to align the Data Center and Cloud road maps, we see opportunities to achieve stronger margins there as well. And we have a few compelling opportunities on the managed services side that we'll talk about in the future. For those deals, we'll adhere to the same core philosophies, but expect strong margins from day 1.

Great. And then on the data center front, it looks like Duke is progressing well on the various phases, at least within the first 99-megawatt delivery and doing so timely. I was wondering if that, along with the sort of potential upsizing to the 300 megawatts for NC-1 is changing the approach or the thought process around acquisition of incremental retrofit sites given that it seems that you might be able to realize additional upside even versus the upside that was contemplated when the asset was acquired. Is that changing how you approach incremental sites, maybe being more open to smaller sites initially given that upside that you might be able to realize here?

Billy, Data Center, do you want to answer that?

So it's not really changing our approach, Paul. I mean that is our approach. We look for sites that have a good amount of power there on day 1, like that we can cash flow quickly and then move on to another project, execute a Phase 1 there and then come back to the initial project and do a Phase 2 or Phase 3. So it's pretty much a copy of what we're doing in North Carolina. We knew that there was a path forward to getting more power at the site. We just didn't know how much and the timing around it. But part of our due diligence and the discussions that we have with utility providers, it's always trying to gauge, yes, here's what's there on day 1, but what can we do going forward. So our approach hasn't really changed, and that is our approach.

Just maybe just to clarify also on that, are you contributing anything in the sense of hardware or other services to facilitate Duke delivering on and upsizing beyond the 200 that was maybe the original view now to 300?

We're still in the early phases of that and studying that with their engineers and our engineers. But yes, there's some land on the property that we need to give them so they can build a new substation.

[Operator Instructions] We'll go next to George Sutton with Craig-Hallum.

This is Logan on for George. So I mean, you've talked quite a bit today about execution being the constraint, not demand. And I was wondering if you could just frame that up for us a bit. Is there a way to kind of size the capacity that the team would be capable of handling from a construction standpoint?

And as we think about there maybe being two more deals coming, it sounds like in the relative near term, are those sites where you've already started to order some of the longer lead time equipment. I just want to get a better sense for like what the size of the constraint is here in a demand environment that certainly sounds pretty good.

Go ahead.

Yes. So the actual construction, project management, manpower, boots on the ground, we're able to easily scale that because of the processes and procedures that we have put in place. And like you outlined, it's really supply chain issues is being able to locate that sheer quantity of equipment to deliver these types of projects. What we can say right now, I mean, our thesis, our approach to this is North Carolina copy pasted 2x or 3x over.

Okay. Got it. And then just a quick one on the financing side for me. I'm curious, as you move to that potentially being done closer to when NC-1 is actually going to be live and generating cash flow. Does that help improve the terms in any way relative to if you had completed financing, say, during the construction phase where there was maybe more risk involved with delivery?

I mean that's pretty self-intuitive. I think you could almost answer that question yourself. But Erke, do you want to give some thoughts around that?

Yes, absolutely. Yes, we do see increasing engagement from our discussions with vendors and counterparties as we're approaching the complete the construction and it's definitely more favorable for securing better terms for financing.

We'll go next to the line of Brian Dobson with Clear Street.

So we're seeing rising demand across the sector for compute data centers. Do you think that, call it, over the next year or so, is that going to translate into better contract terms? Or do you think just the contract signing momentum might pick up? I guess, how do you think this rising demand environment will play out for you?

Billy, data center, do you want to take that?

I think it will play out for us really well because we've put ourselves in a position where we can execute much faster than our peers. So in this capacity-constrained environment, that's really important.

And on top of that, we also have a proven track record of operating. That's -- I feel like that is an important point that often gets miss-looked in the industry. Sometimes people are building these types of facilities and only thinking about the operations towards the tail-end of their projects. And these sites are so complex, and there's so many moving parts and different types of technologies involved that the operation of these large-scale facilities really requires a strong team with experience, and we have that coupled with the speed that we bring these sites online puts us in a really good position in what is a competitive market right now.

Yes. I want to add a few words about that. I'd like to remind all listeners that Billy and his team, they've been working on the retrofit format speed-to-market formula over the past 15 years now, I think even more than that. And there's a lot of team continuity in that time line. And so they were in this sector well before how the sector is today. Back then, it was a boring sector and no one really talked about it. Now it's -- there's a lot of focus and chatter about this sector.

But Billy and his team, one of the reasons why we acquired Enovum and Billy is the CEO of Enovum, which we acquired a couple of years ago, was really because of their retrofit skill set and the storied track record they have in doing this retrofit format for a very, very long time, and they've even done it for the likes of hyperscalers like Microsoft and Amazon. And they already have a great track record in doing it under the WhiteFiber banner, including Cerebras and others. NC-1 is our fourth facility that we acquired that's about to come online for Nscale.

And so I agree with Billy. There's a lot of people who seem to forget that we have a very specialized team on the retrofit side. And that is very important because speed to market is why we have such a pregnant pipeline of demand and customers who are banging on our doors and asking to engage with our team because we can get things up and running very quickly, unlike greenfield projects.

Yes, that's great. In fact, that was my follow-up question about the two potential clients you discussed. Do you think that, that's what drew them into negotiations with you?

Absolutely. But Billy, feel free.

For sure, it's one of the factors. I mean they have orders in for equipment, GPUs, processing power for 2027 that they need to find a home for. And I mean that's one of the gating factors, existing relationships, existing clients, all looking to grow with us.

[Operator Instructions] We'll go next to the line of Sherif Elmaghrabi with BTIG.

Are you in a position to market for NC-1, are you in a position to market any more of the 54 megawatts of capacity? Or is it better to wait for more capacity at the site?

Billy, do you want to take that?

Yes. The initial 54 megawatts is what we're calling Phase 1, and that's fully contracted with Nscale. We have an additional 45 megawatts that is set to deliver towards middle -- between the middle and the end of 2027, and we'll be marketing that right after we deliver Phase 1 at NC-1. And important to note, I mean, obviously, there's no -- we would be offering that first way of capacity to Nscale, just they are our preferred client there, but we would still have the opportunity to market that space if we wanted to.

Yes, that's helpful. And then zooming out, as you guys talked about on the call, power and power equipment is in -- there's a lot of demand. So is there anything you can do to avoid a tight supply chain impacting future projects? Or am I putting the horse before the car here?

I mean it's a very important part of our planning process. Like North Carolina 1, there was equipment ordered before we had clients signed, stuff that's site agnostic, generators, UPSs, stuff that does have longer lead time. So preplanning and having all that prepared to integrate into our project time lines is a really important part. And again, our experienced suppliers that we've worked with for over a decade now, all of that helps us improve and sharpen those time lines.

Our next question comes from Michael Donovan with Compass Point.

I was hoping you could further discuss your strategy around power. So for potential site acquisitions, are you prioritizing those that have power from interconnected utilities like Duke Energy?

Yes. Everything we're looking at right now is grid power. We're not looking at any self-generation or off-grid type solutions. We feel that those are -- first, they're extremely complicated to operate. CapEx for those really complicates the overall profitability of the project. So we prefer on-grid utility-type power at every site we're looking at.

Great. Appreciate that. And then for the medium voltage switchgear, is that switchgear for the entire 40 megawatts? Is it a portion of it? How should we think about the delivery for Nscale would be a little bit lumpy or just hoping to get a little more color on that on revenue contribution and timing.

So it really is -- it's a fluid -- it's still a little bit fluid. It really is fresh. It's a couple of days. That's news of the last couple of days. Initially, it was the entire lineup. It has since improved. So we will be delivering capacity on a scaled ramp-up type delivery instead of a one lump sum on day 1. And that also -- it also helps our clients in testing and commissioning this stuff as well.

So yes, we've gone through a handful of supply chain-related issues in the last 45 days or so, mitigated most of them. Most of them fell into our buffer allocation that we had for the project. And this one is just a couple of days fresh, and we're working on it through our equipment suppliers, our general contractor and directly with the client.

Yes. Just also mentioning that there is nothing material that's being -- the delay here is not material and capacity will begin over the next few weeks.

I think another important point to consider is that the time line that we're executing on, if this was an 18-month delivery, which is pretty much the standard that all our peers are working with, this delay would have not even -- it would have fallen into larger buffer zones or it would have gone unnoticed because we're delivering this in 6 months. This -- any slight delay will have impact. But like Sam said, we don't expect it to be anything material.

We'll return to Nick Giles with B. Riley Securities.

I just wanted to ask about CapEx. Q1 is always going to be kind of a larger spend quarter. Should we expect a similar magnitude in 2Q as you wrap up NC-1? Or how could CapEx kind of progress throughout the year from the sites within your portfolio today?

I think, Erke, would you like to take that?

Sure. At this moment, it's a little bit hard to just provide a very precise number in terms of CapEx spend. And I just break down from data center and Cloud Services. For data center, after NC-1, it will be other sites we'll be looking at. And once that's been finalized, we'll certainly start spending CapEx on that. And on the cost side, Billy mentioned as well, one of the criteria for our contract is that is not going to take a material balance sheet from the corporate. It will be a procurement based on the customer prepayments and GPU banks project level financing. So again, those are all deal by deal. So once the deal being announced, you'll see more clarity. But it's basically a deal-by-deal case.

Okay. Understood. And then maybe one more for Billy K, if I could. We're seeing varying lease structures emerge. And so I was curious if your thinking around lease structures for future deals has changed, maybe how you pass through certain costs to customers and how that might impact the margin profile of any future deals?

We're not really looking to change our structure right now. The only thing that we have been seeing is higher NRCs and setup fees due to custom or more complex type of deliveries that would go outside the scope of basic data center services. But on the margins and the monthly costs, we don't -- we have been approached with certain different types of structures, but we're not looking to change our billing profile right now.

At this time, we have no further questions. I'd like to turn the floor back to our speakers for any additional or closing remarks.

I was on mute. My apologies. Thank you, everyone, for joining us today. We look forward to speaking with you in the next quarter. That will be one that we are definitely looking forward to engaging with you all. Thank you so much for today.

This concludes today's conference. We thank you for your participation. You may disconnect your lines at this time.

Hello, and welcome to the WhiteFiber Fourth Quarter 2025 Earnings Conference Call. Thank you for joining us. We'll begin with prepared remarks from management followed by a question-and-answer session. [Operator Instructions] As a reminder, today's conference is being recorded.

I would now like to turn the call over to your host, Cameron Schnier, Senior Vice President of Capital Markets and Corporate Strategy at WhiteFiber. Cameron, please go ahead.

Thank you, and welcome to the WhiteFiber Fourth Quarter 2025 Earnings Call. Joining me today are Sam Tabar, our Chief Executive Officer; and Erke Huang, our Chief Financial Officer.

Before we begin, I'd like to remind everyone that some of the statements we make on this call are forward-looking in nature and subject to risks and uncertainties that could cause actual results to differ materially. Such risks and uncertainties include, but are not limited to, those factors described in today's earnings press release, our Form 10-K for the year ended December 31, 2025, filed today as well as our other filings we make with the SEC from time to time. Our remarks today may also include non-GAAP financial measures. Reconciliations to the most directly comparable GAAP measures can be found in our Form 10-K and in the earnings press release posted on our website. Following our prepared remarks, we'll open the line for questions.

With that, I'll turn the call over to Sam to discuss our performance. Sam?

Thank you, Cam, and thank you for joining us. Today, we will review our 2025 results and key operational developments. 2025 was a pivotal year for WhiteFiber. We completed our IPO and began operating as a stand-alone public company while continuing to advance and scale our infrastructure platform. As part of that transition, we established the reporting and governance framework required to operate independently. Operationally, the year was defined by execution and positioning for the next phase of growth.

In the fourth quarter, we brought Montreal-3 online. We converted an existing industrial factory into a custom-built data center in approximately six months. That time line reflects our retrofit-first strategy. It allows us to deliver capacity faster than traditional ground-up development. It accelerates time to revenue and reduces development risk. In a market where speed to power matters, we believe that is a meaningful advantage.

Montreal-3 was delivered in support of Cerebras, a leading AI infrastructure company known for its high-performance inference and training systems. We are proud to support their growth. We look to partner with high-quality operators that are scaling quickly. Our goal is to support them as long-term infrastructure partners as their requirements grow.

We are in the process of exercising our purchase option for Montreal-3 for approximately CAD 24 million, funded through our Royal Bank of Canada facility. This is expected to reduce lease payments by approximately CAD 3.1 million annually over the remaining term.

Another major milestone in the fourth quarter was signing a contract for the initial capacity at NC-1. In December, we announced a 40-megawatt IT load agreement with Nscale. The contract carries a 10-year term and represents approximately $865 million of contracted revenue, inclusive of escalators and installation-related services. It covers 40 megawatts of critical IT load delivered in two phases of 20 megawatts each. Billing for the initial phase was originally targeted to begin at the end of April 2026 with the second phase expected to follow one month later. More recently, the customer requested certain design modifications through a change order, which shifted the ready-for-service date for the first tranche by one month. Now the full 40-megawatt IT load is scheduled for a May 31 RFS date. Importantly, these changes are customer-driven and the associated costs are covered through nonrecurring charges under the contract.

Construction continues to progress as planned. Core infrastructure is well advanced, and we are focused on completing the remaining work to bring both phases online. Vendor delivery dates are on time. We have a fabrication shop on site, which accelerates production and improves visibility. Generators are on hand and the remaining work is primarily fit-out, which is underway. Overall, the project is materially derisked.

It's important to note that we didn't take control of NC-1 until late May of 2025. The site was a former textile manufacturing facility, and we didn't begin preconstruction work until the third quarter. Throughout that process, we were in advanced discussions with several potential counterparties regarding a colocation agreement. Ultimately, we engaged with Nscale. We first began discussions with Nscale in October. And by December, we had a fully executed agreement in place. Moving from first discussion to a signed agreement in roughly two months is unusual for a project of this scale. It reflects strong operational and philosophical alignment between our teams. Nscale has an exceptionally experienced data center team that understands the complexity of rapid deployment while maintaining high build standards. Since signing, Nscale has made meaningful progress advancing the commercial structure around this project. Most notably, they have now executed an offtake agreement with an investment-grade hyperscale customer supporting the deployment.

This is an important milestone for the project. The presence of an investment-grade end customer materially strengthens the credit profile of the contracted cash flows. This is expected to expand the universe of potential financing partners, improve overall financing flexibility and support a lower cost of capital as we move forward with the NC-1 financing process. One takeaway from that process is that we are now more deliberate in how we signal the timing of customer announcements. These contracts involve multiple parallel work streams, including technical diligence, site validation, financing and broader commercial structuring, which do not always move at the same pace.

Going through NC-1 helped us refine our approach. Today, we focus on bringing agreements to market once the key commercial and capital structure elements are aligned. This ultimately positions projects to move quickly from announcement to execution. The speed at which NC-1 progressed from acquisition to contracted deployment highlights one of WhiteFiber's core strengths. Our ability to execute quickly is driven by the experience and quality of our team. We continue to evaluate more than 1 gigawatt of power across our development pipeline.

Customer demand continues to exceed our current ability to build at the scale and breadth requested. We are seeing a wide range of demand. This includes smaller 5 to 20-megawatt deployments in specific urban locations as well as multi-hundred megawatt campuses where geography is more flexible. We are well positioned to serve both ends of that spectrum.

Not every deployment needs to be a large-scale campus to be attractive from a returns perspective. In many cases, smaller deployments can be brought online more quickly and are well suited to enterprise customers with more localized or on-premise infrastructure needs. Many of the opportunities we're evaluating today involve hyperscale or investment-grade counterparties and increasingly include structured commercial frameworks that support long-term contracted deployments. These customers require more extensive diligence and longer time lines, but they ultimately result in higher quality, more durable contracts.

Our planning process remains customer-oriented. We respond to customer needs both in real time and proactively as we evaluate new sites. At the same time, we're focused on optimizing our customer base. We are taking a disciplined approach to capital allocation and prioritizing deployments with high-quality counterparties. This includes hyperscale customers and those supported by strong underlying credit that drive durable revenue streams and attractive unit economics.

This has extended the diligence process for site selection. Hyperscale customers are very specific technical -- has very specific technical and operational requirements, and there are many details that must be verified to ensure that a site is suitable. That said, we are well advanced in diligence on several opportunities. One site, in particular, is strongly aligned with a specific -- with specific customer requirements that we're seeking today. The site has strong power and connectivity and is well suited for a range of AI workloads, including latency-sensitive inference. We are completing final technical and commercial diligence to confirm alignment and determine whether we move forward.

We're also seeing increasingly -- we are also increasingly seeing inbound inquiries from landowners and power developers who are seeking experienced partners to develop AI infrastructure on their sites. These partnership opportunities could allow us to expand the platform in a capital-efficient way while leveraging third-party power and land. Our focus remains on expanding the platform, bringing new sites online quickly and demonstrating the flywheel effect that comes from delivering reliable infrastructure for the most demanding workloads. Speed to market remains the defining competitive advantage of our colocation platform and of WhiteFiber more broadly.

I'll conclude the data center section with a brief update on Montreal-2. Montreal-2 is a smaller site in our portfolio, and our thinking around how to best use the asset has evolved as the broader platform has developed. We continue to see two potential paths for the site. One is positioning it to support a hyperscale or enterprise customer and using it as a launch pad for a broader relationship across the platform. The other is pursuing alternate structures that will allow us to optimize capital allocation and potentially redeploy the capital into larger scale opportunities. We are engaged in discussions with a number of high-quality counterparties, and we are evaluating multiple potential structures for this asset. We remain focused on maximizing value as we determine the best path forward within the broader platform.

So let me close the data center portion by outlining our key near-term priorities. First, our focus is on successfully bringing NC-1 online and moving the facility towards stable operations. Second, we are working towards establishing a long-term financing structure as the project moves towards stabilization. Third, we are working towards crystallizing the next tranche of capacity at NC-1 beyond the initial 40 billable megawatts and bringing that capacity to market. We expect to have greater visibility on timing around midyear and would begin marketing these additional megawatts at that point.

Nscale maintains certain rights with respect to the next tranche of capacity and any future allocation would be considered in light of overall platform objectives, including capital considerations and financing conditions. More broadly, we continue to advance discussions around expanding total power capacity at the site beyond our current agreement as well as evaluating potential behind-the-meter solutions to support longer-term growth.

And finally, we remain focused on advancing the next site within our development pipeline. Our objective is to bring at least one additional site and customer deployment forward during 2026. As we pursue that next deployment, we will remain disciplined in optimizing our customer base. We will prioritize hyperscale and enterprise opportunities that best align with the long-term economics of our platform.

Turning now to our Cloud segment. As the GPU cloud market continues to evolve, we are increasingly focused on enterprise deployments and managed infrastructure services rather than commodity bare metal leasing. Our objective is to deploy GPU infrastructure where it complements our broader data center platform and leverages our networking and orchestration opportunity capabilities.

Capital discipline remains central in how we allocate resources. We're not pursuing cloud revenue growth at the expense of funding the expansion of our colocation platform. Since year-end, we have taken several steps to reposition the cloud business. We strengthened our go-to-market organization with leadership focused on enterprise customers. We refined our commercial strategy towards larger and longer-term duration deployments with higher-quality counterparties. And we also began developing partnerships with select Neo clouds to expand reach and accelerate deal flow.

As part of this shift, we monetized approximately 1,000 H200 GPUs for approximately $26 million at a price close to cost. This allows us to redeploy capital into newer generation infrastructure supporting larger, longer-duration enterprise deployments. In addition, our first customer decided to transition away from its contracts as part of a shift towards more flexible cloud consumption. Under those agreements, the customer is obligated to pay a termination fee of approximately 40% of the remaining contract value. We have already redeployed that capacity into new agreements with existing counterparties, including a two-year contract with approximately $50 million in total revenue -- in total value. Together with prepayments, this allows us to recycle capital into new enterprise deployments.

We have also placed our B200 and GB200 capacity across multiple counterparties, representing approximately $13 million of annualized revenue with a focus on longer duration contracts to improve visibility. As a result, approximately 80% of our monthly recurring revenue is now under contract with a weighted average remaining duration of about 22 months. This is a meaningful shift from a year ago and aligns with our go-forward strategy.

Our pro forma fleet consists of approximately 3,700 GPUs across multiple NVIDIA architectures. The majority are deployed under contract with a smaller portion reserved for R&D and future enterprise deployments. As we implement this strategy, we expect cloud revenue to decline in the first half of 2026. This is driven by the hardware lead times and longer ramp cycles for enterprise deployments. We expect approximately $16 million to $17 million of revenue in the first quarter with April representing the low point. Based on our pipeline, we expect revenue to begin ramping in mid-Q2 and accelerate through the year.

We are advancing several enterprise deployments supported by next-generation hardware. As we convert these opportunities, we expect a transition toward more durable revenue streams beginning in the second quarter with the potential for a meaningful ramp in the second half of the year. We expect margins to remain relatively consistent despite the near-term revenue transition as certain fixed costs scale down alongside revenue. We're not pursuing cloud revenue at any cost. We are focused on high-quality customer are focused on high-quality customers, strong contract structures and deployments that meet our return thresholds.

Beyond GPU capacity, we continue to invest in technologies that differentiate this platform, including high-performance networking and distributed training capabilities across multiple sites. Over time, we expect these capabilities to support a greater mix of enterprise deployments and managed infrastructure services while expanding into licensed technology. Our goal is not to operate the largest GPU fleet, but to build a cloud platform that creates durable long-term value alongside our data center business.

I'll now pass the line to Erke to discuss our financial results.

Thank you, Sam. I will review our fourth quarter results and then discuss the balance sheet. Fourth quarter revenue was $23.6 million. This compares to $20.2 million in the third quarter and $14.6 million in the fourth quarter of 2026. Cloud services revenue was $19.3 million, up from $18 million in the prior quarter. Colocation revenue was $3.9 million, up from $1.7 million in the third quarter. The increase was mainly due to MTL-3 coming online during the quarter and the start of revenue from our service colocation contract. This was only a partial quarter of revenue from that contract. Cost of revenue increased as new capacity came online. Gross margin, excluding depreciation, improved to approximately 61% compared to approximately 52% in the fourth quarter of 2026. Depreciation was $8.1 million, up from $6.4 million in the prior quarter, mainly due to the MTL-3 facility. General and administrative expense was $11.4 million. This compares to $21.3 million in the third quarter, which included higher costs related to becoming a public company. We expect first quarter G&A to be slightly higher than the fourth quarter, primarily reflecting increased headcount and the ongoing platform expansion. We expect those investments to support growth as we move through the year. Operating loss for the quarter was $5.4 million compared to $14.5 million in the prior quarter. Net loss for the quarter was $1.5 million. Adjusted EBITDA for the quarter was $5.8 million, representing an adjusted EBITDA margin of 25%. For the full year, adjusted EBITDA was $17.3 million.

Turning to the balance sheet. We ended the year with $114.4 million of cash and cash equivalents and no funded debt. We also had $3.9 million restricted cash and undrawn credit facility with the Royal Bank of Canada. In January, we completed a $230 million convertible note offering due in 2031 with a 4.5% coupon. The initial conversion price is $25.91 per share, which was about 27% above the share price and pricing. We also entered into a zero strike call structure, which raises effective conversion price to about $37 per share and reduces the potential dilution from the notes. The remaining proceeds will support data center expansion and infrastructure investments. We believe our balance sheet and liquidity positions us well to support continued growth.

I will pass the line back to Sam for closing remarks.

Thanks, Erke. Before we move to Q&A, I'd like to spend a few minutes discussing the financing side. Securing cost-effective debt financing for NC-1 remains one of our top priorities. The process has taken longer than initially expected, and we now anticipate putting debt financing in place during the second quarter of 2026.

As we've progressed, lenders have placed increasing emphasis on the overall quality and durability of contracted cash flows supporting these assets. This has made underwriting more rigorous and extended time lines, but also reflect the importance of having fully structured commercial frameworks in place. We believe this ultimately supports higher quality financing outcomes as projects move towards stabilization. At the same time, NC-1 is moving closer to completion and stabilization and the overall credit profile supporting the asset has strengthened meaningfully as the commercial structure around the project has progressed. This includes Nscale's recently executed offtake with an investment-grade hyperscale customer.

We believe this development is particularly important as it enhances the durability and quality of the contracted cash flows and expands the universe of financing partners able to underwrite the project. As a result, we believe we now have greater flexibility in structuring the capital stack and expect improved financing terms relative to where we were earlier in the process. While this has extended the time line, it positions the project to be financed on terms that better reflect the quality of the asset and its customer relationships.

From a capital standpoint, we have clear visibility into the remaining spend of completing NC-1. The majority of core infrastructure is already in place and much of the remaining spend relates to just the fit out. In parallel, we are advancing several financing initiatives. This includes discussions to upsize and amend our existing Royal Bank of Canada facility to support U.S. development activity. We also have access to additional nondilutive capital, including bridge financing. We are actively engaged with financing partners as part of our capital planning. This allows us to maintain a strong liquidity position and continue advancing the platform while we work toward a more permanent financing solution. Our recent convertible offering was primarily intended to support future growth. It also provides additional flexibility as we move through the final stages of NC-1 construction.

Taken together with our existing liquidity, we believe we are well positioned to fully fund NC-1 through completion and retain flexibility to advance the next phase of our development pipeline. This process is also shaping how we approach future developments. We are working with customers and financing partners to ensure projects are structured to be efficiently financed from the outset, which we believe will streamline execution as we scale.

In summary, we're confident in our ability to fully fund and complete NC-1 with the resources available today. The project is currently funded with equity, and our focus is on securing efficient debt financing to recycle capital and accelerate our development flywheel. We're still in the early stages of building a scaled AI infrastructure platform. We have demonstrated the ability to secure high-quality contracts and execute complex developments. Our focus now is on delivering NC-1 and expanding the platform in a disciplined way. In parallel, we are actively advancing our next project and evaluating more streamlined approaches to financing and delivery, reflecting the level of demand that we're seeing and positioning us to accelerate deployment across the platform. We're excited about the opportunities ahead and the role that we play in supporting the next wave of AI infrastructure demand.

Thank you. I'll now open the line for questions. As a note, we have WhiteFiber's President, Billy Krassakopoulos, who will be on the line for Q&A.

[Operator Instructions] We will take our first question from Darren Aftahi with ROTH Capital.

Congrats on all the success. Just two, if I may. Sam, you made some comments that I think a few things have to happen. And again, I might be paraphrasing your words in order for you guys to kind of move forward with additional power on NC-1 and then kind of there's some moving pieces. I guess, can you characterize what are those kind of dominoes that need to fall in order for that to happen?

And then my second question is, you talked about smaller tranches of power, 5 to 20 and then larger scale sort of triple digit. Can you kind of characterize the customers that are looking at both those tranches?

Yes. Darren, just to clarify the first question, are you referring to the additional tranche of power available between -- on NC-1 in connection to Duke Energy? Is that what you're referring to?

Correct.

Okay. Billy, do you want to share with Darren the latest and greatest on that?

Sure. Hi, Darren. So, at NC-1, we've got a commitment to serve from Duke Energy that guarantees us that power. And we're just going through the process with them on the delivery and what milestones need to happen on their side, equipment delivery, planning of their schedule for the upgrade of our substation, which is on our property. So just going through the steps with them on getting that additional power that they -- again, they do have a commitment to serve for that.

And the second question is how we select clients or the clients that are interested in the smaller sites. Is that color on the clients that are interested in smaller sites. Is that correct?

Well, I'm just going to get kind of an umbrella understanding. You're talking about sites that are 5 to 20 and then it sounds like you're hinting that you're looking at other things that are maybe 100-plus megawatts and in conversations with customers that may fit that profile. I guess what I'm just trying to ask is, is it really the same customer base that's looking at all this? Or is it a different profile for each kind of smaller tranche versus larger sites?

Billy, do you want to take that one as well?

Yes. So it's a similar base. Even the larger -- the upper end of the spectrum, even some hyperscalers are looking at smaller deployments in very specialized and urban markets. And through our experience through what we've done in Montreal, we've convinced them that we can turn around these projects quite quickly for them. So that really ups their interest in this capacity-constrained environment.

And maybe a little bit more color on the selection of counterparties. We are laser-focused on minimizing counterparty risk, obviously, maximizing growth prospects and ultimately signing customers that we can finance cost effectively to maximize returns for equity holders. That's our path forward.

We will take our next question from John Todaro with Needham & Company.

Congrats on the progress here. Two for me. I guess the first one on the change orders, certainly seeing that from some others as well. I guess just if we could drill a little bit more into what ultimately drives that? Is that kind of latest NVIDIA architecture that's coming out that the customer wants to slate in? Does this result in any kind of CapEx change? Does anything adjust from that on your guys' end?

Yes. Billy is on the front lines to that. Billy, do you want to take that?

Yes. Hi, John. So the change order is really driven by our clients offtaker on just some optionality that they wanted in how they deliver their networking. It does incur some CapEx changes, but that's primarily passed on to our clients directly.

Okay. Understood...

It has nothing to do with NVIDIA architecture. It's really just more of a physical layout of the end user space.

Okay. Understood. That's very helpful. And then just as it relates to getting the financing done on the back of that offtake that is coming in, Typically, I guess, we see these as like four to five years. Is that sort of what Nscale has lined up to? That's more the time frame we should be thinking of versus -- it wouldn't be something like a 10-year that would cover the full duration of your lease, right? We should be thinking more like four to five years for that offtaker agreement for Nscale.

I'll jump in there, John. That's not really something that we would want or in a position to directionally disclose. So I think you could look at market precedent, but it's not something we're in ability to really give great detail on right now. Sorry, go ahead.

Yes. So we're not -- we don't want to show our hand because I'm sure they're listening.

We will take our next question from Brian Dobson with Clear Street LLC.

So you had mentioned that a client is canceling services and will pay a breakup fee. I guess how would you characterize the current demand environment? And how soon do you think you'll be able to replace that business?

Yes. I mean our initial customer -- we're talking about the cloud side, right, obviously.

Yes.

Our initial customer on the cloud side elected to transition away from the H100 contract, which is set to conclude at the end of this month. That capacity has already been recontracted to an existing customer under a two-year agreement with a one-year extension option with revenue starting in mid-April. So the total contract value there is about $50 million over the initial term. On the B200 side, that contract was terminated in mid-Q1, and we've redeployed that capacity across two counterparties. One is a two-year agreement at roughly $8.4 million annually. And the other is a one-year agreement at about $3 million annually.

So while the annual revenue is lower in the longer term -- in the near term, rather, the replacement contracts extend duration and increase total contracted revenue relative to what remains under the original agreement. So that improves visibility even if it comes with a near-term unfortunate reset.

Yes, very good. And then you mentioned potentially announcing a new facility this year. Would that be a retrofit facility or something more similar to NC-1? What are you looking for?

Yes. NC-1 is a retrofit. And just taking a step back about our retrofit approach, the team that we have was a team that we acquired about 1.5 years ago. They've been doing retrofit models for the past 15 years for the likes of Amazon and Microsoft. And so that skill set was one of the main reasons why we wanted to acquire that team. Billy is the CEO of that team. And so we look at facilities, and I'm speaking for Billy, he's on the call, so I'll stop in a moment, and he could discuss more. But we mainly look for facilities that -- like, for example, in Canada, it was once upon a time a mattress factory. We converted that into a Tier 3 data center within months. on time, within budget. That client, Cerebras is extremely happy with us on that. And for Nscale, NC-1 is built on time. And it's just something -- the retrofit approach is just something we're really, really good at because there's a particular formula and particular -- it's a box within a box model that Billy and his team have basically perfected over the many years now. So we're really happy to get that skill set.

I know there are a lot of peers who are getting into the data center business and doing it from greenfield. We think greenfield, we're not against it. We know how to do it. but we just think there's a lot of execution risk with greenfield, whereas with the retrofit model and the way Billy does it, we can -- the execution risk, the development risk is highly mitigated and we can get things up and running in six months in about 40% cheaper. So this is one of the reasons why there are a lot of customers who are approaching us. They want things up and running really quickly. We're now developing a reputation for that posture that we have. And so we're always looking for facilities like the textile factory in North Carolina or the mattress factory in Canada and others.

We're always looking for facilities like that, that have to have particular bones and there's a whole checklist that Billy goes through that maybe you want to talk about, Billy. But it really shows off the experience and the skill set, and that's why we're getting a lot of particular business because of that reputation that's currently surfacing about WhiteFiber.

Billy, do you want to talk a little bit more. Go ahead, please go ahead.

I misspoke. I meant in terms of scale, but thanks for the additional color on the retrofit, I appreciate it.

I see. I see. Sorry, do you want to ask the question again and maybe Billy can answer it more directly. I misunderstood then.

No, no, I misspoke. I meant in terms of scale, what -- which one of your existing locations would -- could this new project perhaps be more similar to?

I see. Billy, do you want to take that?

Yes. We've got a couple of options. It would be similar to the North Carolina project.

Wonderful.

That is wonderful indeed. We're very excited about it.

We will take our next question from George Sutton with Craig-Hallum.

First, it's great to hear that Nscale has an offtaker for the first leg of NC-1. As we look at the building out NC-1, given the fact that they have a hyperscaler there, does that make them more likely to be the taker of additional capacity at NC-1?

Yes. They are likely to -- I mean, we can't -- we expect that we'll be in a position to market the next tranche of power in NC-1 around midyear. It's likely to correspond shortly after NC-1 is online and drawing power. But Nscale has a prior notification for the site, and we think there is a pretty good chance that they'll want to contract that power.

Got you. Relative to customer 1 transitioning away from B100 and -- I'm sorry, H100 and B200s, the market for certainly the spot side of this has really exploded higher, which is really nice to see. I'm curious, as you've been negotiating and working on these new relationships, are you benefiting from that? Do you feel -- because I guess you said you've extended duration and you've increased the contract sizes. But in the short term, there's a little bit of an impact. So is this really just a function of choosing a longer duration contract at a lower price than taking advantage of the higher spot markets? Just wanted to clarify that.

Yes. We expect the first half of the year to reflect these contract transitions. So revenue will be softer. From a cadence perspective, we're expecting roughly about 15% to 20% sequential decline in the first quarter, with April currently shaping up at the low point of what's contracted today. So, beyond that, it becomes more dependent on the timing of enterprise deployments we're actively working on. If one of the larger opportunities we're pursuing closes, the timing of that deployment could shift to the second quarter outcome and drive a more meaningful ramp in the second half of the year.

More broadly, the focus is on building a more durable enterprise customer base rather than maximizing short-term utilization. Our base expectation for cloud revenue is to be substantially higher in the second half of this year relative to the first half. But again, we'll only pursue growth on terms that make sense holistically for the company.

If I could just sneak one other in. You had mentioned the potential of doing some nondilutive bridge financing. I'm not really clear what you mean by that. Are you referring to just a bridge debt facility? Is that what you're suggesting there?

Yes, Erke, do you want to take that?

Yes. We're working with a few counterparties for getting a bridge to ensure we have the liquidity and to build out NC-1 and then we're looking to get that done in the next few weeks.

We will take our next question from Paul Golding with Macquarie Capital.

Congrats on all the progress. I wanted to ask first as a housekeeping question just on Billy's comments regarding the substation upgrade discussions with Duke. Is that a project that you would be funding yourself or that you could fund into to accelerate that capacity coming online sooner?

And then I have a follow-up around the retrofit environment. I'll just throw that question in now, I guess. As we think about the retrofit model being so core to your competitive edge, how is the landscape for those available projects looking as I don't know if maybe the market is catching on to that being a viable model or even a preferred model. Are you finding that site availability and pricing is still available to you at levels prior? Or is the market getting a bit pricier and getting a bit more constrained around retrofit?

Yes, those are good questions. Billy, do you want to take that in reverse order?

Yes. So, site availability, we haven't seen any changes in that. Site pricing has gone up a little bit. Property owners see the news and see this boom going on in the AI industry. And the news about power and the shortages of power. And so pricing has ticked up a little bit, but site availability, we still have in our pipeline more sites than we can actually enact on.

For Duke and the delivery of the extra megawatts at NC-1, it's really a question about equipment availability and timing. The utility companies have certain schedules where they are allowed to take down certain portions of their network to do upgrades and/or maintenance. So it's really balancing those two factors with the utility company.

Billy, I guess as we think about the constraints on the power infrastructure itself and long lead times, is that something you're looking to mitigate for maybe prospective sites by acquiring some of that hardware ahead of time? How should we think about that since it seems to be a bottleneck you're seeing in the marketplace?

Yes. So similar to what we're doing in North Carolina, we reserve production slots in the -- in certain manufacturers' pipelines that we know have long lead time items. Most of the stuff is site agnostic. It can be placed at any location. So we reserve our slots and -- on the utility side, part of our due diligence, it's a similar model to North Carolina. Part of our due diligence is to make sure that there's a good amount of power there on day 1 with kind of a quick and easy button to upgrading down the road.

We will take our next question from Michael Donovan with Compass Point.

So the K mentioned lease capacity in Atlanta, can you share what the scale is there and what applications you're targeting? Would this be colo or for cloud compute?

Billy, do you want to take that? Sorry, I was...

Yes. If I may add, it's a cloud R&D project. So we're deploying some of the servers for testing and validating our cross data center workloads work stream. And we've contracted a couple of data centers in that region and deploying as we speak, other servers we can test out R&D. And once the R&D is done, we're looking to sell those capacities to compute particularly to end customers.

Appreciate that, Erke. And then one additional one, if I may. Sam, you mentioned behind the need of power. I was hoping you could expand upon that and plans around fuel cells. Is this mainly about speeding time to revenue or expanding beyond current utility applications?

Yes, Billy, do you want to take that.

Allocations.

Yes. Michael, so it's a bit of both. Some projects or client prospects that we looked at were a little short on power available from the utility company. So we look to supplement with natural gas for a temporary period until we get an upgrade on the utility side. We've also looked at the fuel cell solution, which it's a little more pricey, a little bit more of a long-term solution than natural gas. So again, balancing all these with availability of power in certain projects versus customer demand is what's driving.

And just going back to your earlier question, our cross data center R&D project is really well underway, as you mentioned, in reference to the data centers in Atlanta. This -- the impact of that innovation is multifaceted. We think it will be transformative. A promise, this technology that we're working on promises to deliver on colocation solutions to customers where we have a super virtual super cluster -- a virtual super cluster. And if we can nail that technology down, we'll be able to license it to others.

So the two clusters that we're currently connecting in Atlanta are connected across an 85-kilometer connection through dark fiber. These -- the clusters are operational, and we'll begin testing this in the coming weeks. So we plan to report results in two phases starting in April. And the patent filing process is aggressively underway, and the project has cleared screens from prior art. So we're really excited about that. And it's something we're just looking forward to fleshing out and commercializing before we provide a more comprehensive update.

Great. Appreciate the color. Congrats on the progress.

We will take our next question from Kevin Dede with H.C. Wainwright.

I'm curious, Sam could you kind of categorize how you might see turnover at MTL-1 given, I guess, what's happened with customer one, which I'm assuming is Iceland. Is there -- are there any ramifications there?

Can you just say that one more time? I didn't quite catch you on your question.

Sure. I'm curious about customer turnover given the changes that you've seen with customer 1. Does that -- are there implications with MTL-1 and your customer base there?

No. I mean the way we're thinking about the cloud business, I think you're asking about whether the cloud business is shrinking structurally. It's not necessarily. In the near term, we expect a resumption of growth as the business transitions through the repositioning I was talking about earlier. And what we're doing here is we're shifting the cloud segment towards enterprise deployments and managed infrastructure services. So that's where we believe we'll have a stronger competitive advantage. That includes focusing on longer term, longer duration contracts and higher-quality customers. And so as a result, we're just being a little bit more selective on how we deploy capital.

The objective, again, is not to maximize revenue growth in the near term. We believe that we do have the key elements for a successful enterprise-focused cloud business. That said, enterprise deployments do involve longer sales cycles. So there's going to be naturally some timing like there's some timing mismatch. Over time, we do think that we do expect demand for the dedicated and hybrid infrastructure to expand meaningfully, and we're very well positioned to serve that demand.

And Kevin, just to be clear, there's absolutely no bearing on that cloud customer to the colo side of Montreal.

Yes, Kevin, that doesn't affect Montreal at all. And turnover at Montreal-1 right now is not an issue. It's a very highly connected site, urban location that currently is air cooled, but can be adopted for water cooled direct-to chip solutions. So no bearing on Montreal-1 there.

I didn't hear that correctly. It was in relation to Montreal. There's no connection between Montreal and that business.

Okay. Sam, could you give us a little color on your Cerebras relationship? Obviously, you said that they're happy with the development that you've provided at MTL-3. I'm just curious about how -- yes, how do you see their view on the OpenAI deal that they've struck for -- I mean, it was obviously a much bigger facility. And I'm curious if you're running the CS3 technology at MTL-3 and that was maybe a proving ground for them?

Billy, do you want to take that?

Yes. So personally, I'm very extremely bullish on the Cerebras relationship. What we can't disclose the exact model of the system they're running. But from as far as I know, it's the latest version of their technology. And the Open AI deal was a huge announcement on their behalf. And we're looking at expanding that relationship with Cerebras and helping them to serve other markets and other projects as well.

Okay. Last one is just a question on the Duke Energy upgrade for NC-1. Is that a requirement for the second half of the, I guess, 50 megawatts deployment there?

No. The initial deployment is fully secured from Duke. The power is actually being laid up in the next 15 to 25 days to be able to serve the second tranche of the Nscale contract. The power that we spoke about earlier in the call is really for 2020 -- late 2027 type delivery. So...

Okay. So that would be the...

Yes, the full Nscale contract will have enough power from Duke Energy by the time the RFS date.

Right. But so just to clarify for me, Billy, sorry, I thought NC-1 had a total capacity of about 100 megawatts all in, and this upgrade is required for the second half of that 100 megawatts?

That's correct. But the Nscale contract requires about 54 megawatts of gross power to deliver the 40 contracted. So that's all in place from Duke Energy. There's no upgrade required for that. The upgrade is to get from 54 to the 100.

We will take our next question from Nick Giles with B. Riley Securities.

A lot of detail already. But Sam, I think you made a comment in your prepared remarks that you want to be more deliberate with customer announcements. And just hoping you could clarify what that looks like. Does this imply that you may not make any announcement until the data center customer has an end user fully secured? Is this more related to financing? What does that look like?

Yes, exactly. It's -- there are different work streams involved in making all this happen. And either we feed drip we just think that in the next announcement with clients, we just want to make sure everything is the financing and so on is all set up because otherwise, it becomes some occasionally a torturous process, I think, between the capital markets and what we announced. So I just think that the next time, it will be more -- we're going to -- we just want the financing structure in place as we make a mention of the end customer. So that's our posture. It can change..

Cam, do you have any thoughts on that?

Well, we would also just like to be more reserved with respect to how we signal or hint at future contract wins. We don't want to -- we want to just be a little bit more coy and not necessarily see any negotiation ground with customers and keep a few different stakeholders happy through that process.

Got it. Got it. That's helpful, guys. Maybe just one more. I think you heard -- I heard you say that marketing of a new site would occur at the time of the announcement. But it sounds like you're also having several streams of discussions today with offtakers looking for a specific site. So which one would you say is the main priority? Or are these opportunities really working in parallel?

I think they work in parallel but priority to hyperscalers, of course.

Understood. Keep it up.

There are no further questions at this time. I will turn the conference back to Mr. Tabar for any additional or closing remarks.

Thank you. Thank you for joining us today. We look forward to -- we're thankful for your continued support, and we look forward to the next quarter. We are working very hard in executing everything we've talked about today. Thank you very much.

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

Hello, and welcome to the WhiteFiber Third Quarter 2025 Earnings Conference Call. Good afternoon, and thank you for joining us. We will begin with prepared remarks from management, followed by a question-and-answer session. [Operator Instructions] As a reminder, today's conference is being recorded.

I would now like to turn the call over to your host, Cameron Schnier, Senior Vice President of Capital Markets and Corporate Strategy at WhiteFiber. Cameron, please go ahead.

Thank you, and welcome to the WhiteFiber Third Quarter 2025 Earnings Call. Joining me today are Sam Tabar, our Chief Executive Officer; and Erke Huang, our Chief Financial Officer.

Before we begin, I'd like to remind everyone that some of the statements we make on this call are forward-looking in nature and subject to risks and uncertainties that could cause actual results to differ materially. Such risks and uncertainties include, but are not limited to, those factors described in today's earnings press release, our Form 10-Q for the quarter ended September 30, 2025, filed today, as well as our other filings we may make with the SEC from time to time.

Our remarks today may also include non-GAAP financial measures. Reconciliations to the most directly comparable GAAP measures can be found in our Form 10-Q and in the earnings press release posted on our website.

Following our prepared remarks, we will open the call for questions.

With that, I'll turn the call over to Sam to discuss our performance. Sam?

Thank you, Cam, and thank you all for joining us. The third quarter marked the WhiteFiber's first reporting period following our IPO in August. And therefore, it represents only a partial quarter as a stand-alone public company. This period reflects the natural lumpiness of a business in transition as we align operations and capital structure for our next stage of growth.

Our focus remains on building long-term value by executing on our development pipeline, expanding customer relationships and advancing our financing initiatives. I will start with an update on NC-1, our flagship development project in North Carolina. The site remains our top priority. Design, utility interconnection and preconstruction work are advancing on schedule. We remain on track for initial deployments in the first quarter of 2026. We expect NC-1 to begin generating revenue in May of 2026.

Over the past several months, we have seen a significant increase in demand for credible, near-term, high-density capacity. NC-1 is one of the very few sites in North America with meaningful availability in early 2026. That scarcity has made it a highly sought-after resource. As we expanded commercial engagements around the site, we received more than 10 firm proposals from leading AI labs and enterprise customers.

Commercial terms for this delivery window has also improved meaningfully, both in pricing and in overall duration. We are now in the closing stages of negotiations with multiple highly creditworthy counterparties. These are large-scale, long-duration, billion-dollar commitments for our customers. By enterprise standards, the process has progressed at a rapid pace. We continue to work collaboratively through final diligence tied to project readiness and delivery timing. We remain confident in the path to execution.

As a young company operating in a market that has seen its share of unproven entrants, credibility matters. Customers are making generational infrastructure decisions. Our responsibility is to demonstrate readiness, reliability and the ability to execute. We focus on delivering what we promise. We will not compromise on counterparty strength or deal structure simply to announce a transaction. Taking a disciplined approach now positions us to secure that partnership that maximizes long-term equity value and provides clear visibility on future cash flows.

In parallel, we are working closely with our debt advisers, credit partners and prospective lenders to structure the agreement that will be ultimately finalized to be readily financeable on cost-effective terms. We expect the facility to target roughly 75% loan-to-value with the remaining equity portion funded through existing liquidity, operating cash flow or other available sources. Our goal is to secure financing that supports the full 99-megawatt campus.

We also want to maintain balance sheet flexibility and align capital deployment with contracted demand. In short, NC-1 remains firmly on track. We believe the discipline that we've shown has strengthened both the commercial and financial profile of the project. We remain focused on execution and on building a cornerstone asset that will drive significant value for years to come. We believe the discipline and the patience that we've shown are now yielding a better result. We expect to finalize an anchor agreement in the very, very, very near term.

The depth of demand we saw for NC-1 also reinforces how broad the opportunity set is for WhiteFiber. Several of the counterparties that engage with us on NC-1 have expressed interest in capacity in other regions where we do not yet have assets. Based on those conversations, we are narrowing in on a new development site that offers meaningful 2026 power availability and strong expansion potential. Client demand is guiding our site selection process. We are now actively progressing through diligence.

Overall, our development pipeline remains very strong. We are currently evaluating more than 1 gigawatt of projects for the 2026 and 2027 timeframes. Our first priority remains finalizing the NC-1 anchor and moving that project through financing and construction. The interest we have seen underscores the scale of opportunity ahead.

The actionable portion of our pipeline is growing rapidly. We're driving that growth by identifying and securing high-quality sites that others often overlook or can't access through traditional channels. This capability has become a key differentiator for WhiteFiber, as we build a platform for sustained growth.

Turning to Montreal. Our Montreal-3 facility is now operational and began recognizing revenue from our contract with Cerebras in October of this year. The retrofit was completed on schedule and within budget. For the fourth quarter, we expect Montreal-3 to contribute slightly more than $2 million of revenue with a full quarter contribution beginning in the first quarter of 2026.

Revenue for this site should be approximately $1 million per month depending on the Canadian-USD exchange rate. Completing this complex retrofit in under 6 months, from site control to revenue generation is a testament to the speed and precision of our development and operations team. That level of execution has not gone unnoticed by customers and partners. We are already receiving inbound inquiries to replicate this model at larger scale. Reliable delivery is often the best form of marketing in our industry.

At Montreal-2, we continue to evaluate both colocation and internal cloud use cases. Activation has become -- has been deferred while we prioritize NC-1. The facility remains a flexible asset that can be deployed quickly once the commercial path is finalized.

Turning to Cloud. The market narrative has shifted. Early last year, supply was extremely tight. Since then, supply chains have improved. New capital has also funded additional entrants. As a result, the tone has become more skeptical. This change has created pockets of price-led competition. Some providers are treating GPUs as a commodity. Our view is simple. Capacity that competes on price alone without matching performance, reliability or software capability will struggle to attract long-term financing.

Over time, the market will reward operators that deliver real outcomes, not just low sticker prices. In that environment, our approach remains patient and selective. Our growing colocation business supports the broader platform. Because of that, we do not need to chase uneconomic growth. We only expand when contracts meet our return thresholds. That includes multiyear terms, upfront commitments and take-or-pay structures.

We are also comfortable waiting through periods of market softness when that is the right call. When the economics are attractive, we face a different constraint. The challenge is often data center space within the customer's delivery window. Larger cluster opportunities appear regularly. The limiting factor is usually the availability of rack space at the required time.

Today, we run our GPU cloud business from third-party facilities. We believe our capital is better used developing and owning data centers for colocation. The risk-reward profile is stronger there than building data centers solely for our own GPU clusters. Capital is finite, and we are disciplined on how we deploy it. We stay flexible, but we remain strict about returns. We have walked away from deals that did not meet our standards. That discipline has served us well.

We continue to evaluate a wide range of opportunities. We are focused on deals with at least a 12-month term and pricing aligned with our target payback period. Recent smaller wins that have met those criteria have consumed much of our previously available inventory. Since quarter end, we began winding down a customer agreement. It represents a little over $20 million of our annualized cloud run rate.

Discussions are ongoing, so we can't share details or identify the counterparty just yet. We expect a mutual termination once the documentation is complete. Even so, we believe redirecting this capacity to larger enterprise and institutional customers is the right long-term strategy. It also helps us avoid potential complications.

We have already identified new counterparties to absorb this capacity. The commercial terms are comparable or frankly, even better. These contracts should begin contributing in early first quarter. Because of this transition, we expect a short term of underutilization in the fourth quarter. We still plan to partially monetize the capacity through on-demand pools. That should limit any temporary revenue impact.

Looking ahead, we are exploring more capital-light models in cloud. Those include managed services, partnerships and enterprise private cloud deployments. These approaches use our software and operational strengths. They also expand our addressable market, while offering attractive returns on invested capital.

We are also investing in technology that improves the long-term profile of the business. Our recent engineering hires are enhancing our software stack. They are improving deployment flexibility and enabling cluster performance that exceeds standard benchmarks. These investments help us compete on performance and reliability, not price. They support durable growth across the cloud platform.

Overall, our Cloud platform remains a recurring high-margin business. It complements our colocation and development activities. As remarketing progresses and new contracts are signed, we expect to restore ARR and grow through the coming quarters.

And now, I'd like to pass the line to Erke to discuss our financial results. Erke?

Thank you, Sam. Total revenue for the third quarter was $20.2 million, up 64% year-over-year from $12.3 million in the same period of 2024. Cloud services generated $18 million of revenue, an increase of 48% from $12.2 million a year ago. The growth was driven by the expansion of GPU capacity for new and existing customers, partially offset by a $2 million service credit recognized this quarter under a customer agreement.

Colocation services contributed $1.7 million of revenue compared to none in the prior year period following the integration of our Enovum operations. We also recorded a $0.5 million of other revenue from equipment leasing activities, bringing total consolidated revenue to $20.2 million.

Total gross profit was $12.7 million, representing a gross margin of approximately 63%. Cloud services accounted for $11.7 million of segment gross profit and colocation services for $1 million. Operating expenses were $34.7 million compared to $13.1 million a year ago. General and administrative expense was $21.3 million, which included approximately $11.3 million of non-cash share-based compensation.

The share-based compensation was elevated this quarter due to the one-time initial grants for both existing employees and key hires. We expect this expense to be significantly lower in the fourth quarter in the low single-digit million range. We estimate our quarterly run rate core G&A to be around $6 million on a cash basis. In addition, there's around $1 million to $2 million overhead associated with new data centers coming online, project level costs tied to closing data center sites or negotiating customer contracts and certain discretionary items, such as events or incremental marketing spend.

There were also a number of one-time or nonrecurring items in the third quarter, including costs related to public company readiness, deposits on data center sites, recruiting expenses as we expanded headcount and other miscellaneous items. We have expanded our organization to support a significantly higher revenue base, which should drive operating leverage as we scale. Several recent hires will also allow us to replace consulting expenses as those functions transition in-house.

Excluding nonrecurring items, normalized G&A -- cash G&A would have been approximately $8 million for the third quarter. We expect the fourth quarter G&A to be around $10 million, including share-based compensation. Operating loss for the quarter was $14.5 million compared to a loss of $0.8 million in the third quarter of 2024. Net loss was $15.8 million or $0.47 per share compared to a net loss of $0.4 million or $0.01 per share in the period -- same period last year. On a non-GAAP basis, adjusted EBITDA was $2.3 million compared to $5.6 million in the third quarter of 2024.

Turning to balance sheet. We ended the quarter with $166.5 million of cash and cash equivalents, up from $11.7 million at the year-end 2024, reflecting the IPO proceeds received in August. Capital expenditures totaled $13 million for the quarter, driven primarily by the completion of Montreal-3 and initial spend for North Carolina-1.

We expect 4Q CapEx to increase materially based on budgeted work for the development of NC-1 and potential discretionary spend on GPU procurement. Potential GPU spend will be predicted on contracts rather than speculative inventory build. We remain a very strong liquidity position with working capital of approximately $179 million and no draws on our existing credit facility with the Royal Bank of Canada.

I will now turn the call back to Sam for closing remarks.

Thank you, Erke. As we close today's call, I wanted to take a step back. WhiteFiber is still early in its journey as a stand-alone public company. This quarter marks our transition from launch to execution. Our current cost base reflects the upfront investments required to operate as a public company. It also reflects the cost of supporting a much larger business. We have built the foundation to scale.

As revenue grows, we expect meaningfully operating leverage to follow. Our focus remains exactly where it should be. We are earning the trust of our customers, partners and shareholders by executing with discipline. We are building a business designed to endure beyond cycles. Our strategy is simple, develop and own data centers that generate long-lived cash flow, grow a high-margin cloud platform that scales responsibly and finance both in a way that protects shareholder value. NC-1 will be the cornerstone of that platform.

The commercial process is now in its very final stages. We believe the discipline we have shown will lead to a stronger, long-term outcome for our shareholders. At the same time, the level of enterprise demand we continue to see gives us confidence. It tells us we're pursuing the right opportunities. It also confirms that high-quality capacity with near-term availability is increasingly scarce.

Our priorities for the coming quarters are clear: finalize the NC anchor, secure financing for the full campus, and continued scaling colocation and cloud in a balanced, capital-efficient way. The opportunity in front of us is significant. We are positioned to execute with credibility and speed. I want to thank our employees for their focus, our partners for their collaboration and our shareholders for their continued support.

With that, operator, please open the line for questions. As a note, Ben Lamson, Head of Revenue, and Billy Krassakopoulos, President of WhiteFiber, will be presented for Q&A. Please open the line.

[Operator Instructions] We'll now go to your first question that will come from the line of Darren Aftahi with ROTH.

Congrats on the progress. I guess, as you talk about the developmental power pipeline, I just had a question on that. Are you broadly sticking to kind of this overlooked asset strategy where you can kind of repurpose/retrofit sites and properties? Or is the strategy much broader than that?

And then my second question is, I assume you've gotten inbounds, and you kind of hinted that at some other sites. Are you taking more of a strategy where you're trying to procure sites and power on behalf of customers? Or is this procure the site and then figure out the formal lease with a potential customer after the fact?

Yes, those are great questions. Easy to answer. Billy, do you want to take those questions?

Sure. So, Darren, we're doing both. We've got projects that are retrofits. We've got projects that we're looking at that are greenfields, and it's really customer interest and customer demand that will drive final decisions on those.

Second part of your question, again, we're doing both. We've got properties that are specialized for specific situations, specific clients, depending on their timeline. We also have properties that we're looking at that we are taking the build it, and they will come approach. But again, very conservative and strategic in making those decisions.

We'll now go to your next question coming from the line of Nick Giles with B. Riley Securities.

I was hoping you could provide some additional color on why the customer agreement at NC-1 is taking a little longer. Is it really driven by additional parties coming to the table, maybe due diligence taking longer than expected? Or is it really down -- still down to ongoing negotiations on terms? Just curious at this point, how many parties you remain in discussions with for that initial capacity?

Yes. The answer -- what was the first part? Because the answer was yes to number one, yes to two and no to three. But can you remind me of the first question? There were 3 questions there.

Yes, sorry. Sorry, Sam.

Rather, there are 3 permutations. So if you could just remind me the permutations.

Yes. Maybe just at this stage, how many parties are you in discussions with? Are you really down to one party? And then -- yes.

Yes. So we had a tsunami of interest from very high-quality counterparties, counterparties you've heard of. We reduced it to 2. There are now 2 who are competing extremely hard in finalizing this. These agreements, by the way, required multiple steps. They include engineering work, commercial negotiations, internal approvals. These counterparties are giant, so they have to go through their own Board approvals. So that naturally extended the timeline a little bit. But we have very firm proposals in hand. I was expecting to sign one today. But now, it's just a little bit of due diligence and confirmatory due diligence on both parties.

They're both equal in terms of attractiveness. I will be countersigning the one who signs first. But one factor behind the longer timeline is there were a lot of parties involved. Additional groups, as you suggested, entered the process during the quarter. The demand continued to build. And by the way, that did put upward pressure on both pricing and term length. So that was great to have.

We also worked closely with our debt advisers. We wanted to avoid rushing into a structure that would be harder or more expensive to finance. So we wanted to be disciplined on that there. So that took a little bit longer given the multiple counterparties we're speaking with.

So yes, we've been patient. We've been deliberate. We believe that's the right approach. We're now very close to the goal line. And as of this morning, I thought I was going to be signing one of the agreements just before this call, those very final steps did not align in time for this call. So we'll provide an update as soon as both sides finish their processes.

I appreciate all that detail. My follow-up would be, how much activity is going on at the site today? Are you -- it sounds like you've reiterated your timelines for May, for the initial phase. So are you still continuing with development or if things really at a stage where you're pausing until you sign that first customer?

Yes. I know the answer to that, but I'd love for Billy to take on that because he's on the front line.

Sure. So as of now, we're not -- we haven't paused anything. We're still continuing a lot of the demolition and the cleanup work of the site completed a couple of weeks ago. Municipal permitting and stuff like that is all in process. And nothing -- to date, nothing has slowed down.

Next question will come from the line of Brian Dobson with Clear Street.

Very positive news regarding the imminent contract signing. You mentioned that you were in negotiations with like 10 potential clients at the onset. And then, of course, that winnows down to a few. Are you seeing this similarly -- yes, are you seeing similarly robust demand at your other facilities? And are you seeing similarly robust demand for, call it, yet to be contemplated facilities that are being proposed by these clients?

The answer is yes. And Billy, do you want to take that on?

Sure. Our other facilities do not have the volume that North Carolina had. So the counterparties, "runner-ups" that we had for this property, are looking at our portfolio and asking what kind of availabilities we have for end of 2026, early 2027. So a lot of positive stuff came out from us just analyzing all the demand that we had and not signing with the first party that came to the table.

Also, probably, Billy, worth mentioning -- is it worth mentioning that a lot of these clients as we have been speaking with them are looking to future sites that we've identified?

That's right. They're looking in our portfolio. They're looking at our pipeline and trying to match what fits in their schedules of deployment. So like we said, we've gotten a lot of interest. The runners-up in the NC deals are looking to do projects with us in late 2026, early 2027.

And then, in the past, you've said that you would contemplate both large-scale and smaller-scale facilities. Do you still feel that way? And do you still see, call it, more attractive return on investment at the smaller ones, even if marginally so?

The environment that we're in, the shortage of inventory and capacity that we're currently going through, gives priority to both. We've got smaller sites that are 30 megawatts that we can execute much quicker on that are in the pipeline with attached customer demand. We also have 100-plus megawatt sites, again, with attached customer demand that we're looking and currently due diligencing and analyzing to execute on.

Next question will come from the line of George Sutton with Craig-Hallum.

First, just a quick question on North Carolina. So I think you're suggesting that either of the 2 customers would take the full site, is that correct?

Billy, do you want to take that since you're interacting directly with them?

That is correct.

Okay. On the GPU as a service market...

Probably worth mentioning the deal -- I know I mentioned this already, but I'll say it again, the deal did upgrade as the negotiations went on. So time was our friend. And so we've upgraded it in terms of profit and duration. So it's been a good problem to have. But of course, the negotiations went on a little bit longer as the deal upgraded.

Understand. So relative to the GPU as a Service market, Ben, I wondered if you could just talk more broadly about the pricing you're seeing, the contract duration you're seeing, the demand by the NVIDIA generations, what is different about that market today than perhaps a quarter ago? And then separately, can you just give us any details in terms of the timing of the wind down of the customer agreement you mentioned?

Yes. So to address your first couple -- there's a few things going back there. So I'm going to make sure I touch on all of it. Still extremely strong demand for H200, even H100, in fact. I just had an inquiry for H100s come through today. So still really strong demand for Hopper generation. B200s -- B300s are starting to get delivered.

I think what we're seeing across the market is there -- some of our peers are taking large deals at price points that we just don't feel makes sense, and we feel that they're irresponsible. And to us, we feel that some of our peers are taking these deals to just to be able to announce a logo. When in reality, it's not financially responsible. And we're -- on the other hand, we're practicing pretty extreme financial discipline. And I think that's going to help us weather any upcoming storm.

So yes, there's pricing pressure in the industry, but I think it's predominantly driven by some, what we call, irresponsible decisions that we're not going to participate in, where we're focused on longer-term agreements with healthy margins, where we are competing on performance and reliability and not competing on price. And in the long run, we're pretty confident that that's going to pay off.

Did I unpack -- was there anything else I missed in that kind of first half?

No, I think that's good. I am curious if you're getting any benefit from some of the kind of first-to-market technologies that you're offering. Is that yet in market? And is that part of your discussions?

Yes. Yes, yes. So yes, we absolutely are getting some benefit there, not yet in market, but it has generated quite a bit of buzz and partnership inquiries. And we think once we get this white paper out early next year in Q1, it's going to open a lot of doors. I mean, there's already a lot of conversations happening behind the scenes right now that are early. Far, far too early to talk about, but we're certainly really excited and optimistic for what things like the cross data center workload are going to open up for us. Yes, I hope that answers your questions.

That does. Can you address the timing of the contract that is getting wound down just so we have some sense of the impacts?

Yes, that's still ongoing. That hasn't been completely finalized. So I don't have details that I can share on this call at this time.

Yes. We're still in the process of discussing the mutual termination.

Next question will come from the line of Paul Golding with Macquarie Capital.

Congrats on the progress. I wanted to ask, Sam, just to clarify, we're referencing these deals are for anchoring NC-1. Is that specific to the first 24 megawatts in Phase 1? Does it extend to Phase 2? I guess, how far into the 99 initial megawatts are these conversations or your expectation in terms of this deal being signed? And then I have a follow-up.

Billy, do you want to take that?

Yes. So all of the conversations we're having extend to the Phase 2 of delivery, which is just a little north of 15 gross megawatts. So it's essentially doubled from what we were looking at last, and that is pretty much all that Duke Energy will be able to supply us in 2026. The remainder of the 99 comes online in 2027. So -- and it attests to the strength and the types of customers that we're seeing. So the deal basically upsized and doubled within the last couple of weeks/months.

That's great. Congratulations on that. I guess, as we think about the potential for '27 to get to that 99, is that second phase something that's already being marketed and that you're getting inbound interest on beyond the counterparty that you've been in discussions with on Phase 1? Or is that something you're holding back for now to see how pricing and availability contribute to those negotiations?

So we call that the third phase of the project, and we've -- all the parties that we're talking to right now would like rights of first refusal on it. So we're just holding that back right now in anticipation of closing this deal.

Next question will come from the line of John Todaro with Needham & Company.

I just want to go back to NC-1 and understand the lease process a little bit better. Were you saying that you -- the customer that you could have initially signed that you could have got that done, but kind of held out for higher pricing or better pricing? What I guess happened to that initial customer? Did they have to bail out? Or are they still one of the two now? Because I thought pricing was already kind of negotiated.

Regarding better pricing and even better duration. Go ahead, Billy, but -- go ahead.

During that process, John, there was a couple of other horses that entered the race to say with longer terms and better pricing. So we naturally had to entertain those. That customer specifically was for lower capacity, and they eventually ended up losing their offtaker. So think about a neocloud that had an offtaker for the capacity, and that's what happened there.

But in the meantime, the deal pretty much -- our lease process pretty much got doubled to 40 megawatts of IT load and a longer deal term than what we initially had on the table. So having to process that and go through all the due diligence and commercial negotiations, internal approvals on both sides is what's really extended the timeline here.

Got it. Understood. That's very helpful, very clear now. And then, just as we do think about the site potentially coming online and generating revenue in May, it does seem soon, and you got some peers out there where there's been some delays in execution misses. Just kind of -- can you frame up the confidence in that, especially because it still might take, I guess, a little bit here to get the lease signed and some of those pieces?

Yes. So I mean, essentially, the train had already left the station a couple of months ago on design and equipment procurement. There is a lot of equipment procurement that's site agnostic that we had already placed on order. There's a couple of minor stuff that really depends on the end user. We're still confident in our dates. Like we said in one of the earlier questions, demolition, site cleanup, early construction work had already started. Applications for permits, construction permits with the local counties had already gone through. So the process is moving forward. The timing is tight, but we're still on track for everything that has been forecasted.

Your next question will be coming from the line of Kevin Dede with H.C. Wainwright.

Sam, would you mind sort of characterizing your position on the GPU business and cloud versus colocation under complete understanding that NC-1 remains your top priority. But it looks like -- and it looks like it will -- I mean, the first 2 phases, right, is a $400 million proposition at least. So -- but I got the sense that you're still interested in supporting that GPU business. I mean, just help me understand your priorities a little bit better, please.

Can you rephrase your question, Kevin? I just want to make sure I understood that.

Yes, sure. No, I understand NC-1 and that development is a priority, but at least through next year, it seems to be a $400 million development proposition. So I'm wondering, with that as the backdrop, how should we think about your prioritization of supporting your own GPU business?

Just to be clear, we're prioritizing the colocation business. That is -- that NC-1 for us is our North Star in securing the anchor, which is imminent, and then our -- and then to execute our financing strategy, which is very straightforward. We plan to finance about 75% of that full project with long-dated asset-backed credit, mats to the duration and stability to the underlying customer contract. That's very important for us. And we've been working very closely with our advisers to ensure that the structure we pursue is correct. And so that's something that we've been all very much focused on.

With respect to the cloud business, we remain very disciplined. There's actually been deals that we could have announced with extremely well-known names, the hottest names in the market, but the economics didn't make complete sense for us. And we didn't want to just announce a sexy headline and have crappy economics when you open the hood, so that's something we steered away from.

So the financing for the first 2 phases of NC-1 are predicated on your counterparty? Is that fair to assume?

There's lots of elements. But yes, it's -- certainly the quality of the counterparty is important. It's one factor for getting good terms on the financing, absolutely. And that's why the counterparty needs to be creditworthy.

So a small site like MTL-2 is in high on your priority list. Would it be worth maybe selling it or somehow leveraging its value and its undeveloped state?

I mean, we've got a really great deal for it. And if we were to sell it, we probably would be able to sell out of profit, frankly. And we've even got -- we even received interest, soft bids for it that were higher than the amount that we purchased it for, but we want to keep it for now. There is still some path towards monetizing that, including some cloud use cases that we'd like to test that site on. We've invested a lot of money on tech, and we'd like to use that particular site to -- for R&D on the cost data center workloads. So that's something that we do not wish to sell just yet, although if we did, we would very likely get quite a handsome profit of selling it.

Right. So regarding the cross data center workload technology, understand that there are other tools in your tool belt, and I was wondering how you see them potentially presenting a competitive differentiation. I mean, there's a big neocloud out there that's been gobbling up other software companies in developing its orchestration stack. And I'm wondering how you see WhiteFiber pair up to that capability.

Well, I mean, good question. Power availability -- and Ben, feel free to add to this, the power availability is one of the biggest bottlenecks in the AI infrastructure ecosystem. We're investing in technology that would allow us to aggregate and orchestrate geographically through distributed power pools, enabling customers to deploy clusters that exceed the limitations of any single substation or a site. This is early-stage technology work, but the direction of travel is clear. Customers want larger clusters, faster than traditional power infrastructure can accommodate. We expect to share more on that on the first quarter of next year.

Ben, this is -- I'm sort of stealing your thunder there, feel free to add.

No, Sam, I think you said it well. I mean, that's going to be a big -- it's a tool. It's a pretty big tool in our tool belt. We are also -- we have homegrown solutions for orchestration that we think are pretty powerful, and we've gotten really good reviews around. And going a layer below that, we're highly focused on pure network performance and as the first cloud to deploy a DriveNets cluster. I mean, the results of -- the benchmarks from that cluster are phenomenal and I think unmatched and something we're really proud of. So I think we are going to continue to develop around performance at the foundational layer of infrastructure and see where that takes us.

So to that point, Ben, is that DriveNets cluster technology something you can develop at MTL-1? Or would you need to fully develop MTL-2 in order to explore it? And with this other contract coming off, what kind of room does that give you to take on a new customer?

We've got a lot of room to take on new customers. I mean, we've got a pipeline right now for that capacity. The DriveNets cluster is different than the cross data center workload, though we are working with DriveNets on that. So I want to make sure that's clear. Those are 2 different things versus a single cluster versus the cross data center workload.

Yes. Understood. I was just wondering if you needed MTL-2 in order to develop that tech.

No. So we've identified a site. Again, we're keeping the cards close to our chest. We have identified sites in the U.S., where we have begun deploying the V1 of our cross data center workload.

Next question will come from the line of Nick Giles with B. Riley Securities.

Maybe just 1 for Billy. Obviously, a lot of time and energy has gone into the initial lease here. And so once that gets done beyond really execution at NC-1, what are your main priorities or goals as we head into 2026? I mean, should we expect to see a new site announcement in the near term? And how much capacity could be on the line?

Sure. Thanks, Nick.

Billy, let's be careful on specifics, of course, so that we don't -- yes.

I mean, obviously, the main focus is North Carolina. These things -- it's a process. These things usually take anywhere from 4 to 6 months just to close. We thought we were able to get it done a bit sooner, but the closing process, as we mentioned in all the other questions, the closing process taking a little longer than expected.

Pipeline, we are continuously looking at and refreshing all the sites in our pipeline trying to align that with customer demand. So looking at new sites with -- in tandem in conjunction with our clients trying to fit sites, projects, timelines is always a priority. It's something that we're continuously working on.

Maybe if I could just add to that. We're -- just like Billy said, we're evaluating a large number of sites. We do it on a rolling basis. And the reason why I don't want to be specific is because the pipeline is really dynamic, sites move in and then move out as we complete diligence. So it's really hard to maintain a single published figure, but today, we are evaluating over 1 gigawatt of potential development opportunities across the U.S. and Canada. These sites vary in timing.

A meaningful subset has power availability in the second half of 2026, while others line up more naturally with 2027 and beyond. The later delivery sites could align very well with our capital deployment cadence, and that depends on how NC-1 progresses in where the customer demand is strongest. But like Billy said, our priority is to fully finalize and derisk NC-1 before we stretch the platform across multiple builds.

That said, through the NC-1 process, we've had customers that have asked us to evaluate specific markets where they want to work with us and be prepared to pre-lease capacity. And these signals, they're more than signals, they're outright instructions almost. They help shape our search. So we're looking at a healthy mix of off-market retrofits and greenfields. There are a few of them that are especially compelling.

Once NC-1 is in a fully derisked position, we expect to shift our attention to selecting and formalizing the next development site. We're not trying to spread capital around until the timing and the commercial visibility is right.

I guess, just, Nick -- I just want to mention one thing related to Nick's question. In terms of the demand for colocation, it's effectively the diametrical opposite of what is playing right now in the capital markets, where the sky seems to be falling on companies linked to the AI ecosystem. We see no shortage of demand.

In fact, it is greater, far greater than it was just a month or 2 ago when the sector was viewed in a shining light. We continue to see greater and broader demand for NC-1 as awareness of the project growth. And we've seen greater demand from new potential customers asking us about sites in different regions to Billy's point. Many customers came to us for NC-1, and then, they began asking us about other sites in specific regions.

So the demand isn't just centered around 2026, but also to 2027 and beyond. And we're seeing similar levels of demand also on the cloud side, but the economics have to be right on the cloud side. So I just wanted to get some further color on that, especially in light of the capital markets today.

Data center infrastructure will always be in demand, and the global shortage in supply will be for us -- well, I think will be there for the considerable future, frankly.

And it appears there are no additional questions at this time. I'll turn it back to you for your closing remarks.

Well, look, thank you very much for your patience today. I wish that we had been able to announce the contract today when we're very close. And it's -- we look forward to making that announcement in the very near-term future.

Thank you, everybody, for your patience, and we'll continue to build out WhiteFiber to the company that it will be in the future. So thank you very much, everybody.

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