White Fiber - Q2 2025
September 17, 2025
Transcript
Operator (participant)
Good afternoon and thank you for joining us. We will begin with prepared remarks from management, followed by a question-and-answer session. During the Q&A, if you would like to ask a question, please press star one on your telephone keypad. 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.
Cameron Schnier (SVP Capital Markets and Corporate Strategy)
Thank you and welcome to the WhiteFiber Second Quarter 2025 Earnings Call. Joining me today are Sam Tabar, our Chief Executive Officer, and Eric Wong, our Chief Financial Officer. Before we begin, I'd like to remind everyone that some of the statements we will 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 fiscal quarter ended June 30, 2025, filed today, as well as our other filings we make with the SEC from time to time. Our remarks today may 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'll open the call up for questions. With that, I'll turn the call over to Sam to discuss our performance. Sam?
Sam Tabar (CEO)
Thank you, Cam, and thank you all for joining us. This is our first earnings call as a standalone public company following the successful completion of our IPO in August. The offering priced on August 6th at $17 per share, the high end of the range, and closed on August 8th. It was upsized due to strong demand. On September 3rd, our underwriters fully exercised the Greenshoe overallotment option, bringing total gross proceeds from the transaction to approximately $183 million. Bit Digital now retains ownership of roughly 71.5% of WhiteFiber. The IPO established WhiteFiber as a pure-play, independent AI infrastructure company. We own, operate, and develop high-performance computing data centers designed specifically for artificial intelligence workloads. One of our differentiators is that we control both layers of the infrastructure stack: the physical data centers and the GPU cloud resources that they host. This integration enables us to develop reliable, cost-efficient, and scalable infrastructure directly to our customers while capturing value at multiple points in the supply chain. As a technical matter, today's results cover the second quarter ended end of June 2025, which were already included in Bit Digital's consolidated financials. Because our IPO occurred in August, this 10-Q was required to bring WhiteFiber current with its SEC reporting obligations. Going forward, we will be on a regular quarterly reporting schedule. While the second quarter results reflect the early stage of our business, the most important development during the quarter was our acquisition of NC1, the data center property in North Carolina. This 1 million sq ft site has significant power availability and expansion potential. It was a pivotal step in scaling our platform. It positions us to meet growing demand from enterprise and hyperscale customers for high-density AI infrastructure. We have commenced the pre-construction phase at NC1 and remain on track to bring the initial 24 MW phase I online in the first quarter of 2026. We are in active discussions with multiple prospective customers, several of whom are seeking capacity beyond phase I, including interest in the full build-out of the site. At this point, we expect the initial contract to cover more megawatts than phase I. One customer recently shifted the base design to NVIDIA's newest hardware. That required a recontracting process, but it's positive; it aligns NC1 with the latest generation of hardware. We continue to progress with this customer and others in parallel. We remain confident that we will be able to formalize customer contracts for at least phase I in the coming weeks. Every week, it seems like we're approached by a larger customer that has serious interest in the site. That said, our priority is speed. We are not willing to draw out the process and allow for extended due diligence for new entrants. All counterparties we are engaged with are creditworthy and are advancing with structures that strengthen the financing profile of the site, including OEM backstops, security deposits, and escrow arrangements. These mechanisms improve our visibility on cash flow and support a lower cost of capital. We have formally launched a process for debt financing at the site, which will allow us to cost-effectively scale the remaining capacity of phase I in line with customer demand. In short, NC1 is a valuable strategic asset, and we are encouraged by the level of customer enthusiasm and demand to date. We continue to expect the capital requirements for building out the site to be approximately $8 million per megawatt. CapEx timing tends to be back-weighted closer to when megawatts are approaching operational status. Turning to MTL-3, we are in the process of bringing the site live and turning on equipment for Cerebras. The first batch of wafer-scale engines arrived on site this week, with additional shipments expected in the coming days. We expect to begin recognizing revenue under the contract in October. This project highlights the execution capability of our team. We signed a 5 MW IT load contract with Cerebras in February of 2025, but we did not take control of the Saint-Jérôme facility, which previously operated as a mattress factory, until late April. In under six months, we have completed a sophisticated custom retrofit and prepared the site for production. That timeline would have been notable for a standard build. For a deployment of this complexity, it is exceptional. We're grateful that Cerebras entrusted us. May someone please mute themselves. Thank you. We are grateful that Cerebras entrusted us with this project and look forward to supporting them for years to come. At MTL-2, the deployment timeline has been adjusted as we prioritize other builds and preserve capital for more time-sensitive projects. Our original plan was to dedicate the site to WhiteFiber's own GPU deployments. Ultimately, we determined that the prospective end customer for that use case did not meet our contract requirements. We have therefore opened the site to potential colocation clients, while keeping the option of a WhiteFiber GPU deployment in the future. We currently expect MTL-2 to be activated in the first half of 2026, although we may accelerate or reprioritize the site based on any future customer developments. Having this operational flexibility is a key strength of WhiteFiber. If we don't have a long-term and economically viable opportunity on the cloud side, we can simply toggle the site for colocation. Ahead, our pipeline remains robust. We will continue to be opportunistic in expansion but disciplined in our capital allocation. With the volume of opportunities in front of us, we're focused on being disciplined in our prioritization to ensure that we execute to the standards we have set for ourselves. As our capital base grows, particularly with the addition of debt financing, we will pursue larger scale opportunities. Our focus remains on driving value at our existing sites while setting the stage for our next deployments. We want to build upon our track record of retrofitting deployments on industry-leading timelines. Customer trust is non-negotiable in AI infrastructure, and we believe our ability to deliver reliably and quickly is a defining strength for WhiteFiber. We also expect to layer in smaller customer-driven sites over the next 6-12 months. In some cases, customers engaged with us at NC1 have expressed interest in capacity in other regions where we do not currently have assets. Expanding selectively in such markets, guided by firm demand by customers, will remain a key element of our strategy. Turning to our cloud business, this remains a key driver of revenue for WhiteFiber. In the second quarter, cloud services generated $16.6 million in revenue, up 33% from the prior year, with gross margins of approximately 61%. Our cloud services run rate revenue is currently approximately $88 million, following the rolloff of some smaller contracts last week. This figure does not include about $3.5 million of additional run rate contract payments from our deployments with Boosteroid, which is reflected in other revenue in our financial statements. The combined run rate is currently over $90 million for our GPU-related business lines. We also hold roughly 700 GPUs that are not yet under contract, including 576 H200s and a mix of B200s and GB200s. At market pricing, these assets represent approximately $15 million of annual revenue potential. In addition, we have an active customer pipeline and are in discussions on several larger opportunities that could commence later this year if awarded. Should we secure these, we would expect to finance the deployment with equipment financing. This is consistent with our approach of aligning capital deployment with specific customer demand rather than building speculative inventory. At the same time, we're continuing to invest in our cloud software platform to deliver highly performant and reliable infrastructure. Combined with our ownership of the underlying data centers, this integration differentiates WhiteFiber from peers that focus solely on GPU cloud. We want customers to choose WhiteFiber because of the performance and reliability of our solutions, not simply excess capacity, and we're continuing to invest in R&D to support this goal. Our product roadmap is focused on delivering GPU clusters that exceed performance benchmarks and customer expectations. As an example, in the second quarter, we were the first to market with scheduled fabric Ethernet technology. This deployment resulted in a 30% performance improvement over industry benchmarks. This focus on technology leadership may also open the door to new capital-like business opportunities over time. Cloud is a key part of WhiteFiber. It generates attractive margins, builds long-term customer relationships, and provides recurring revenue that complements the scaling of our colocation business. I will now hand the line to Eric to discuss our financial results.
Eric Wong (CFO)
Thank you, Sam. Total revenue for the second quarter was $18.7 million, an increase of 48% from $12.6 million in the same period last year. Cloud services generated $16.6 million in revenue, up 33% from $12.5 million a year ago. Gross profit was $10.1 million, representing a margin of about 61% versus $8 million and a 63% margin last year. Colocation services contributed $1.7 million of revenue and $1 million of gross profit, representing a margin of about 60%. We did not generate colocation revenue in the second quarter of 2024. We also recorded $0.3 million of other revenue, primarily from our deployments with Boosteroid, bringing total consolidated revenue to $18.7 million. Total gross profit was $11.5 million, an increase of 43% from $8 million in the same period of 2024. Operating expenses were $27.8 million compared to $10.2 million a year ago. G&A was $15.5 million for the quarter, which included around $5.5 million of one-time milestone stock-based awards and certain non-recurring costs related to the growth of our business and preparation of becoming a public company. We expect normalized G&A to be lower on an ongoing basis, but are not prepared to issue discrete guidance at this time. That said, we have been increasing headcount and preparing the company for a significantly larger operation base. Some of the hires we have made will reduce consulting expenses we have incurred. Operating loss was $9.2 million compared to an operating income of $2.4 million in the same quarter of 2024. Net loss for the quarter was $8.8 million compared to a net income of $1.9 million in the same period last year. On a non-GAAP basis, EBITDA was negative $3.2 million compared to a positive $6.8 million a year ago. After adjusting primarily for share-based compensation expense, adjusted EBITDA was positive $3.3 million versus $7 million in the second quarter of 2024. Turning to the balance sheet, we ended the quarter with $16.4 million of cash and cash equivalents. Property, plants, and equipment grew to $230 million as we expanded our data center portfolio and GPU fleet. Following the IPO, we added approximately $183 million in gross proceeds, which provides us with significant liquidity to support our development pipeline. We also signed a credit facility during the quarter, which remains undrawn. As of June 30th, we carried no debt on our balance sheet. Together, our post-IPO cash position and undrawn facility provide us with a strong liquidity base to fund the initial build-out of NC1 and support cloud growth. I will now hand the line back to Sam for closing remarks.
Sam Tabar (CEO)
Thank you, Eric. Before we open the line for questions, I want to thank our employees, customers, and shareholders for their support as we begin this next chapter as a standalone public company. While today's results reflect a period before our IPO, they highlight the strong foundation we are building across both cloud and colocation. Looking forward, WhiteFiber is uniquely positioned as a pure-play AI infrastructure platform with both the data center and GPU cloud layers under one roof. We believe this integration gives us a competitive advantage and positions us to capture the current wave of enterprise and hyperscale demand. We are early in our journey, but the combination of a growing cloud services platform and rapid progress at sites like NC1 and WhiteFiber 3 demonstrates our ability to execute across the entire infrastructure stack. We look forward to updating you very soon on our continued progress when we report our first full quarter as a public company in November. Our mission is simple: build the best AI infrastructure, build faster and cheaper with our retrofit strategy, meet customers wherever they are, and create lasting value for shareholders. With that, operator, please open the line for questions. As a note, Ben Lampson, Head of Revenue, and Billy Krassakopoulos, President of WhiteFiber, will be present for Q&A. Please open the line.
Operator (participant)
Thank you. If you would like to signal with questions, please press star one on your touch-tone telephone. If you're joining us today using a speakerphone, please make sure your mute function is turned off to allow your signal to reach our equipment. Again, that is star one if you would like to signal with questions. Our first question will come from Darren Aftahi with Roth.
Darren Aftahi (Senior Research Analyst)
Hey, guys, thanks for taking my questions and congrats on the IPO. Two, if I may, just during the roadshow, you guys talked about pipeline power on the colo side and various sites. I'm just kind of curious. Sam, you talked about multiple parties interested in North Carolina. If those other parties don't get that, do you guys have stuff that is more kind of line of sight in that pipeline? My second question, Sam, I know you use the word formalized in the coming weeks on North Carolina. What exactly does that mean? Does that infer, like, a formal lease or just any more color on that would be great? Thanks.
Sam Tabar (CEO)
Yeah, thank you for that. Billy, do you want to take some of those questions? I'm happy to add on.
Billy Krassakopoulos (President)
Sure thing. Thanks, Darren. We have several highly interested parties that we're engaged with on North Carolina. Given their timelines, there isn't much room left to finalize these contracts and still meet their deployment schedules. As a result of that, things are moving quite quickly, and we expect clarity on commitments in the very near term. We're not just waiting for one customer anymore. There are multiple that are interested, and they all understand that they need to move quickly. We can't afford to let any single customer draw out the process when we have others that are ready to move. Some news on what's happened with these interests: one customer, for example, doubled the amount of capacity that they had signed in their initial LOI. Another one pivoted to the latest unreleased NVIDIA hardware. Another one signaled that they would like an additional deployment in another market that we're not even in yet, right after we would deploy for them in North Carolina. It was contingent on a North Carolina deployment. These things take a little bit of time to close, redoing technical plans and financial contracts when the capacity changes on such a high level. Just take a little bit of extra time to close these contracts. All of them are aware their timelines are all Q1 2026, and pen to paper needs to be done quite quickly to accomplish that.
Sam Tabar (CEO)
Greg, any commentary on the aggregate pipeline beyond North Carolina?
Erke Huang (CFO)
We're looking at sites in Arizona, Texas, the state of New York. We are looking at another site in North Carolina as well. Given the timelines, these are still in early stages of development. It would be quite difficult for them to work on any potential client that would not get the first phase or the entirety of North Carolina. The prospects that we have for North Carolina are really all looking for a Q1 2026 delivery.
Sam Tabar (CEO)
It's probably worth mentioning also, Billy, the pipeline of facilities that you're looking at that are not necessarily marketed as data centers. If you want to unpack that a little bit, I think that might be helpful.
Billy Krassakopoulos (President)
Sure. I think the fact that one of the customers that we're speaking with is looking for us to build for them in another market that we didn't even have on our radar after we pitched them the North Carolina project goes to show how much they value our speed in our retrofit model, which is quite important in today's times.
Sam Tabar (CEO)
To add to that, we do build two times faster and 40% cheaper. We have a track record of that. In fact, we're handing over the keys to Cerebras as an example of that in just a few days. We bought that. That was a mattress factory once upon a time. We took that over, I think it was in April, I believe, Billy. Is that correct?
Billy Krassakopoulos (President)
End of April, yeah.
Sam Tabar (CEO)
End of April, and we're able to convert that into a tier three data center for one of the most demanding hardware clients out there in the world, Cerebras. Those keys are being handed over extremely shortly, in days. That's just an example of how we do things very differently from others in this space. We are winning business because of that track record and because of that ability. Time is extremely expensive for these clients. They don't want to wait around for greenfield builds. Billy and his team are able to build these things using the retrofit model while the others are still waiting for the concrete to dry. It's really important for clients. That is why we're getting a lot of this very enthusiastic business interest because of that particular model that we have. I just think it's worth mentioning that.
Darren Aftahi (Senior Research Analyst)
Appreciate it. Thanks, Sam. Thanks, Billy.
Operator (participant)
The next question will come from Nick Giles with B. Riley Securities.
Nick Giles (Senior Research Analyst)
Thank you, operator. Good afternoon, everyone. I just wanted to first follow up on NC1. There seem to be some kind of higher credit quality parties entering the conversations. Do you have a rough deadline in mind of when you'll cut off inquiries? Have you done so already? My main question is really what aspect of the deal is really driving the conversation today. Is this around overall scale, price, timing? Appreciate any additional color there.
Sam Tabar (CEO)
Billy, you want to take that one again? I'll add some context if I feel necessary.
Billy Krassakopoulos (President)
Sure thing. Nick, it's a mix of all of those points. Change in technology on one prospective client's end changes the base of design and changes the base price point. You touched on creditworthiness. A lot of these customers have different grades of creditworthiness in our books. We are not taking on any new inquiries to this site. We are actively engaged with a handful of parties right now and just moving along that process.
Sam Tabar (CEO)
I'll just say it, we're at the horse race between three right now. We're going to optimize for the best. Go ahead, Billy.
Billy Krassakopoulos (President)
We're still aiming to deliver in Q1, end of Q1 2026. There will be an eventual cutoff point where the party that's not as far along as the others will probably have to drop off.
Nick Giles (Senior Research Analyst)
Very helpful. Just one follow-up on that.
Sam Tabar (CEO)
Nick, I think you also asked, there are some who are trying to upsize the megawatt to take the whole thing. If you want to ask a more specific question about that, go ahead.
Nick Giles (Senior Research Analyst)
Exactly, Sam. That was really my follow-up. I think you made some comments around the 24 and the 40. I was wondering in those conversations if you're having any that could include all 99 and if you could really remind us of what would be required from the energy provider to reach that level. Thanks very much.
Sam Tabar (CEO)
Yeah, Billy, you want to take that? I'll add context if necessary.
Billy Krassakopoulos (President)
Sure. Yeah, we did have one party initially contract for 20 and then ask for the full 40, and another one asking for the full 99. There are conversations happening with the utility provider, assurances on timeline for them to deliver for us and for us to deliver to the end user. That's what's really delayed the process on those two specific cases.
Nick Giles (Senior Research Analyst)
Got it. Guys, thanks so much for the update. I'll jump back in the queue, but keep up the good work.
Sam Tabar (CEO)
I just want to also mention you didn't ask, you sort of touched on it, Nick. I just want to give some more information with respect to the quality of the counterparties. They are extremely high. They're creditworthy. Of the three, in my personal opinion, there is one in particular that, I mean, they're all very exciting, but we are just trying to optimize for the best counterparty because that has a cascade effect on credit facility and so on. This is a question of optimizing the best for our situation. This is a question of weeks and not months. I just wanted to give some more color on the tension that's happening. These are happy tensions for us. These are champagne problems. We understand the market is impatient, but we do want to get the very best customer announcement.
Nick Giles (Senior Research Analyst)
Understood, Sam. Thanks again.
Operator (participant)
We'll take a question from Brian Dobson with Clear Street.
Brian Dobson (Senior Research Analyst)
Thanks very much. Good evening and congratulations on the successful IPO. As you're looking at additional data centers, you had mentioned that you might be looking at larger projects or smaller projects. Do you think that you can walk us through what the returns on those types of products might be? Are they roughly similar between large and small? How would you characterize the different types of clients that are looking for, call it large versus small? Are they household names or is there any kind of color you can give us there?
Sam Tabar (CEO)
Billy, you want to take that?
Billy Krassakopoulos (President)
Sure. The household names are usually for the larger projects. We tend to balance where we look at next, mostly through customer demand. The smaller ones are still AI-based, AI users, but deploying more of their own private infrastructure, like the example that I mentioned before for North Carolina. We've got a user speaking for 20 MW of North Carolina, but contingent, we deliver them shortly thereafter another 20 MW in a market that we were not even looking at. It is really based on customer, driven by customer demand and balance of available capital, depending on the size of the project, right?
Sam Tabar (CEO)
Billy, is it fair to say that the smaller sites are juicier in economics, but probably less in stature on the counterparty name? Still well known, but not like elite, but the economics are cheaper with the smaller morsels, as it were, compared to the bigger sites?
Billy Krassakopoulos (President)
Exactly, and a bit of low-hanging fruit in the short term, but the smaller sites, like part of our due diligence process when selecting these properties, are how much power is available now, but a pathway to more power in the future. Even if we were to select a small site right now, it's because the power is available there immediately. One of the checkboxes is, is there a pathway to more power so we can maximize and squeeze that site as much as possible. Thanks.
Brian Dobson (Senior Research Analyst)
Yeah, thanks very much for that clarity. There's been some very exciting news regarding AI data center demand over the past few weeks. As you're looking out through 2026 and 2027, I guess what gets you most excited about growth in the industry?
Billy Krassakopoulos (President)
We think.
Sam Tabar (CEO)
Go ahead, Billy. Go ahead.
Billy Krassakopoulos (President)
This industry, the data center industry specifically, it's always been a build it and they will come business up until about a year, year and a half ago. That's when the flood of requests, significant inbound requests from prospective AI customers, started. I mean, that's really exciting. Finding the right site for the right customer is also an exciting challenge for us.
Sam Tabar (CEO)
I want to add to that. We're slowly getting a reputation in the market for, again, establishing this track record of building 2x faster and 40% cheaper using this very specialized approach that we have. A lot of these facilities that Billy and his team are looking at are not even marketed as data centers. We're able to get these facilities for not the premium that is normally charged when they're marketed as data centers. Once Billy and his team get their hands on these facilities, they're able to convert that into tier three data centers, just like that mattress factory that is now a tier three data center for Cerebras with revenue recognition, I think, happening next month. That's just another example. Billy and his team have been doing this for many, many years before we acquired them last October. That acquisition was very strategic. What excites me is that a lot of people are starting to now understand that greenfield builds have massive execution risk and the retrofit model is only possible through many years of experience, of trial and error, of a proven approach in doing this successfully. People are realizing that WhiteFiber is the answer for that because time is of the essence and time is not the friend of technology. You have to get these technologies online to extract value. They see that the retrofit model being 2x faster with a track record of that, they see that that has a lot of value and they come to us. That excites me about that reputation going around the market. We're seeing that in action already.
Brian Dobson (Senior Research Analyst)
Yeah, thanks. That's great color. You should put 2x faster and 40% cheaper on a hat and wear it around. Thanks a lot.
Sam Tabar (CEO)
That's actually a great idea, Cameron. Let's make sure we get hats for that.
Operator (participant)
Our next question will come from George Sutton with Craig-Hallum.
George Sutton (Senior Research Analyst)
Thank you. I wondered if you could talk about the contract requirements that didn't meet your needs at MTL-2. I mean, are we talking as simple as price, or is there something else to be aware of?
Ben Lampson (Head of Revenue)
Yeah, George, Ben Lampson here. Honestly, go ahead. It was a combination of price and term, but predominantly term. You know, as we've discussed in the roadshow, we're really focused on private cloud deployment and on-prem as a service, and identifying customers that can't and won't put their workloads in the public cloud. For these shorter-term customers or shorter-term contracts, if we wanted to pursue those, we'd really just be pursuing being another public cloud, which is not the arena that we see the biggest opportunity in.
George Sutton (Senior Research Analyst)
Gotcha. Relative to the 700 GPUs that are not under contract, I'm curious, are you marketing that as a single deal? Are you prepared to move to an on-demand service offering if not?
Ben Lampson (Head of Revenue)
Great question. It's a mix of both. They're not in a single cluster; they're across a few different clusters. They're scattered. Some of those are actually getting put on demand. We're excited about some opportunities we have there with some partners. Some of them are, we've got some long-term contracts in the pipeline for a large chunk of those as well. It'll be a blend. We don't suspect that we will have that much free capacity for long.
George Sutton (Senior Research Analyst)
Gotcha. One other question, if I could, for Billy. Hypothetically, if you get this customer that wanted the 20 MW in North Carolina, wants 20 MW somewhere else, can you just talk about what the contract like that, what kind of risks do you have, and do you need to find the site and then develop out the site and go through the regulatory process, or just walk through kind of that dynamic if that ends up being the deal you take?
Billy Krassakopoulos (President)
Sure. We work hand in hand with the customer on selecting the site in this specific scenario. We would do our due diligence and make sure that the site meets all of our requirements. The client would sign off on the site. Once they're contracted, part of our KYC with the client is a really deep credit check, security guarantees, OEM backstops. We wouldn't go into a market that we didn't think we could expand into. The initial phase would be 20 MW. We would work on expansion with whichever utility provider would be in the area and make sure that we can expand and either help this customer grow into that facility or find other customers to fill the remaining capacity that we would get.
George Sutton (Senior Research Analyst)
Gotcha. I understand. Okay, thanks, guys.
Billy Krassakopoulos (President)
Thank you.
Operator (participant)
Thank you. Moving on to Paul Golding with Macquarie Capital.
Paul Golding (Senior Analyst)
Thanks so much. Congratulations on the listing. I wanted to ask on the financing that you are already in conversations on with respect to build-out and CapEx, funding the CapEx. I wanted to ask with the deposit, escrow, and other backing that these counterparties have been offering in these negotiations, could you give us some color around how the cost of capital for project financing is developing midway through these conversations? How much is the creditworthiness and these escrow components and other offerings contributing to those conversations with lenders and delivering a more efficient cost of capital model for your retrofit? I have a follow-up. Thank you.
Erke Huang (CFO)
I can talk about the financing. We have secured a credit facility with RBC, but that's for the MTL site. For the financing process for North Carolina 1, we had engaged the JOL Capital Markets and recently launched a debt raise for securing the financing and build-out of that site. We're still early in the process, given the fluid nature of these discussions, but initial interest has been very strong. More specifically, we have seen significant depth and liquidity in the lending universe. The interest expands in commercial real estate lending markets, insurance company capital, private credit, digital infrastructure funds, and pure play project finance banks. North Carolina is quite unique, given the in-place power of 100 MW, but also near-term expansion of another 100 MW.This provides tenants the opportunity to grow within the facility, and also gives lenders the opportunity to create a relationship with our team and a clear path to additional business. Regarding specific terms, we anticipate financing in a range of 75%-80% loan-to-cost. The debt will find an immediate feed-out, but we are also targeting a future funding facility of recording that we provide WhiteFiber with immediate access to capital to build out additional capacity as the leases are signed. It's too early right now in JOL's process to point on pricing, but we anticipate it to be in line with the recent deals, higher 200 basis points over SOFR or 300 basis points over SOFR. Those are the markets we're looking at.
Paul Golding (Senior Analyst)
Thanks so much for that call. I really appreciate that. Maybe a question for Ben on the cross-data center cloud offering that is part of the productized offering suite of products that you've been discussing with counterparties. How is that product evolving? Could you give us an update on whether there's any early interest if you're marketing that at all or still waiting to further progress in developing that product? Thank you.
Ben Lampson (Head of Revenue)
We are not currently marketing that. It is still very much in the R&D phase. We want to get a little further along with some of the results on that or get closer towards publishing a paper on that before we start to market it. I think that's going to be really, really important from what we've seen in the past, getting that paper out. We have narrowed in on a site for that that we're really excited about, and we're progressing with that process. We're still targeting early next year for the publishing of that paper and for that technology to be live.
Paul Golding (Senior Analyst)
Great. Thanks so much.
Operator (participant)
The next question comes from John Todaro with Needham & Company.
John Todaro (Equity Research Analyst)
Thanks for taking my question and congrats on the progress out of the gates. Just going back to if some of these contracts are upsized, the ones at North Carolina phase I, it sounds like that NC1 phase I, we should expect bigger than the 24 MW. As you think, if the counterparty does sign for the full 99, what would be the timeline for that? I imagine it would be a little bit further out than Q1 2026. I have a follow-up related to that.
Billy Krassakopoulos (President)
Sure thing, Don. Initially, there were 24 MW available at that facility, and 44 MW coming online early Q2 of 2026. We're in talks with Duke Energy to bring that up a little bit forward to meet the initial demand of doubling capacity from 20 MW or so to almost 50 MW. The remaining 99 MW right now is looking like it's about two years out.
John Todaro (Equity Research Analyst)
Got it. Looking two years out. Okay. I believe it was also mentioned a change in the base price point. I would interpret this as potentially better rental economics than some of us were kind of initially expecting and targeting. I believe we were thinking, call it $2 million per net critical megawatt or about $1.6 million, I think, on a gross basis. Is the thinking that you could get even better rental economics than that now?
Billy Krassakopoulos (President)
It is. It's slightly higher, but it could be offset by a higher build cost. Because of this new technology, the build cost would be incrementally higher, but we offset it with an increased price per megawatt.
John Todaro (Equity Research Analyst)
Got it. Understood. You're saying, would we expect incrementally higher than the $8 million per megawatt in the build cost, or does the $8 million include that kind of higher, potentially higher build cost?
Billy Krassakopoulos (President)
It does not include the potentially higher build cost. That is one of the delays with closing one of these LOIs. It is really limited information on the installation specifications of this new hardware from NVIDIA.
John Todaro (Equity Research Analyst)
Oh, understood.
Billy Krassakopoulos (President)
as CEO of WhiteFiber.
John Todaro (Equity Research Analyst)
Sorry, no, please go ahead.
Billy Krassakopoulos (President)
We're basing the new cost and the new financial terms on the contract of the information that we have right now and the build specifications that we're getting from the client.
John Todaro (Equity Research Analyst)
Got it. Okay, that all makes sense. Thank you. Thank you, gentlemen. Appreciate it.
Billy Krassakopoulos (President)
Thank you.
Operator (participant)
The next question will come from Nick Giles with B. Riley Securities.
Nick Giles (Senior Research Analyst)
Thanks so much for taking my follow-up. I just wanted to clarify on MTL-2. Can you just remind us what customers will ultimately be filling that capacity if you look to put your GPUs elsewhere or hold off on incremental contracts? Thanks.
Sam Tabar (CEO)
I'm happy to take that. We originally envisioned MTL-2 to be a flagship site for WhiteFiber GPUs. We advanced discussions with a specific customer over an extended period, but over time, their contractual requirements shifted and no longer really aligned with our underwriting criteria. Things like duration, prepayments, and overall financeability were factors for us. Rather than hold the space while we searched for another GPU fit, we made the decision to open MTL-2 for colocation opportunities. The site could have been operational months ago had we prioritized it, but without the right contractual foundation, it just didn't make sense to us. We are now in discussions with customers on a colocation business while also remaining open to housing a cloud customer there. Our current target to bring that site online is going to be in the first half of 2026, with the option to accelerate depending on demand.
Nick Giles (Senior Research Analyst)
Got it. Thank you very much. I really appreciate that.
Operator (participant)
At this time, there are no further questions. I now turn the conference back over to you.
Sam Tabar (CEO)
Thank you all once again for your time today. It is worth mentioning again that I would like to thank our shareholders for their confidence in us. We are executing, and we look forward to making certain announcements in the very near-term future. Thank you for your trust in us and your patience, and we look forward to next time. Thank you very much, everybody.
Operator (participant)
Thank you. That does conclude today's conference. We do thank you for your participation. Have a nice day.
Speaker 13
Goodbye.