Hello and welcome to the White Fiber Third Quarter 2026 Earnings Conference Call. Good afternoon and thank you for joining us. We will begin with prepared remarks for management, followed by a question and answer session. During the Q&A, if you would like to ask a question, please press star 1 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 White Fiber. Cameron, please go ahead.

Thank you and welcome to the White Fiber Third Quarter 2026 Earnings Call. Joining me today are Sam 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 September 30, 2026, filed today, as well as 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'll 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. We've reached two important milestones at NC1. First, all 40 megawatts of contracted IT capacity are now operational, and we have begun billing for the overall contracted capacity. As a reminder, The contract value is approximately $865 million over 10 years. Second, we've signed a $300 million secured credit facility for the project. Third, since our last earnings call, we've also signed more than $700 million in new cloud services contract value. That includes our first Capital Light managed services agreements. At the same time, Demand for our next North Carolina sites, NC2 and NC3, continues to materially strengthen. Together, these developments give us a stronger foundation for growth in 2027. Our focus now is execution, bringing signed contracts into service, securing the funding to support them, and choosing our next investments carefully.

Let me start with NC1.

Completing this project is a significant milestone. It shows that our retrofit model can deliver complex AI infrastructure at scale. Despite the challenges, we brought 40 megawatts of IT capacity into service in less than 11 months. Getting to full billing took longer than we expected. I want to explain why. Shipping disruptions related to the Middle East conflict delayed our original switchgear. To keep the project moving, we sourced replacement equipment and adapted the electrical design. Our team then had to resolve integration and commissioning issues with that equipment. Billing for the first 20 megawatts haul began in mid-August. A switchgear software issue later interrupted service and billing. Both resumed on September 23rd. Further switchgear commissioning issues delayed the second haul, including the final five megawatts. As a result, we reached full contracted billing later than we expected in August. The issue affecting that final capacity has now been resolved. Full billing for the second 20 megawatt haul has begun on November 9th. All 40 megawatts are now billing under the contract. Importantly, we do not expect late delivery penalties related to these issues. We expect development costs to finish just below $9 million per gross megawatt. That is within our original budget range of $7 to $9 million. That figure excludes the site acquisition costs and is net of customer-funded, non-recurring costs. NC1 gives us an established US operating base and a development model we can build off. We're applying what we learned to procurement, testing, and commissioning to NC2 and NC3. We recently completed the $60 million acquisition of these two sites in Yadkin County. We believe they are exceptionally well suited to our retrofit strategy. They have existing industrial buildings and on-site substation and valuable electrical infrastructure. We expect at least 60 megawatts of initial gross utility capacity. They're also about 55 miles away from NC1. That means we could draw on our existing workforce and regional resources to support development and operations. These sites show why we favor retrofits. Existing buildings and electrical infrastructure gives us a head start. They can shorten delivery timelines and reduce construction and permitting risk compared with building from the ground up. Finding sites with these advantages is one of our strengths. The customer interest we're seeing supports our confidence in the acquisition. We see an opportunity to turn this infrastructure into assets that generate cash for many years. We're targeting initial capacity in late 2027. Both NC2 and NC3 are already zoned for data center use, which removes an important hurdle. Delivery will still depend on power availability, customer commitments, financing, and the remaining approvals. We're also applying our community engagement experience in Madison. Strong local relationships are an important part of how we develop these projects. Across North Carolina, we're in advanced customer discussions for about 105 megawatts of gross capacity delivered for 2027. That includes about 45 megawatts of expansion at NC1 and 60 megawatts across NC2 and 3. Interest in NC2 and 3 is substantially stronger than the interest we saw a year ago when we were marketing the first phase of NC1. We have a broad range of interested customers, more non-binding letters of intent, and stronger prospective counterparties. Some of these proposals come directly from investment grade counterparties. Others include investment grade credit support. That gives us more choice in selecting agreements we can finance and deliver successfully. Demand is also expanding beyond AI infrastructure providers. We're seeing direct interest from Fortune 100 companies for deployments below 50 megawatts, as well as smaller purpose-built projects. Market sentiment has been weak, but the customer demand we see continues to grow. More enterprises want dedicated infrastructure for their own AI programs. We're also seeing opportunities tied to government-related demand. Over time, we expect enterprises and public sector customers to become a larger part of the market. They want dedicated compute, control over sensitive data, and infrastructure located in specific jurisdictions. Our co-location, cloud, and managed service business gives us several ways to meet those needs. Our experience at NC1 has also sharpened our approach to these new projects. We require investment grade credit at the project level, either directly from the customer or through a full credit wrap. We're structuring customer agreements and financing together from the outset. We intend to move quickly toward delivery in 2027, but we will maintain those standards even if that affects the delivery date. As power becomes harder to secure, it matters more whether a site has potential capacity or power that can actually serve a customer. We believe existing electrical infrastructure combined with a credible delivery schedule will become increasingly valuable. Development costs have risen over the past year. Based on our current budgets, however, the pricing and terms and customer proposals would more than offset those increases. That creates an opportunity to earn higher yields on development costs than we achieved at NC1. We also consider financing costs and terms. Ultimately, the return on the capital we commit will guide the agreements that we sign. Looking further ahead, we're preparing to deliver substantially more infrastructure in 2028 and beyond. We're advancing sites and power, strengthening our teams and systems, and developing financing options. We want to be ready for the right opportunities while keeping flexibility when and how much capital we invest. Returning to Canada, MTL2, will host the prime interlock deployment we announced last quarter. The agreement covers 576 Vera Rubin GPUs over the initial three-year term. This will be our first deployment combining our own data center with our cloud services. The budget to complete the facility is about $25 million. Development debt is available through our Royal Bank of Canada facility. were targeting facility readiness in early 2027 ahead of the start of GPU services in the second quarter. Bringing facility design and GPU deployment together gives us more control over delivery and operations. Most of our near-term new data center capacity will continue to serve pure co-location customers. We'll combine our facilities and cloud services selectively where the economics make sense. Owning facilities also gives us more options when GPU contracts expire. Where we have available capacity, we can redeploy equipment to extend its revenue-generating life. We can also serve shorter-term demand when pricing is attractive. At NTL2, we're evaluating behind-the-meter power options that could bring additional capacity online sooner. We're also continuing to pursue utility expansion across our Montreal portfolio. Turning to cloud services, we continue to convert demand into larger multi-year agreements. Since our last earnings call, we've signed approximately $750 million in new contract value across five customers. These contracts cover dedicated GPU services and managed services. They include IG counterparties, new customers, and larger commitments from existing customers. They also establish our cloud presence in Sydney and Dublin. In both cases, customer demand is driving our expansion. The new dedicated compute agreements have an average term of about 4.2 years and cover roughly 13,000 GPUs. That brings our total undersigned dedicated compute contracts to about 20,000 GPUs. The new deployments are scheduled to enter service between late in the fourth quarter and the second quarter of 2027. Once our assigned cloud services portfolio is fully deployed, we expect it to generate about $365 million in annualized revenue. We're targeting that run rate by mid-2027. The timing depends on equipment and data center readiness, customer acceptance, and financing. Based on our current schedule, we expect fourth quarter cloud revenue to be roughly in line with the third quarter's $18.1 million. Growth in ongoing services should offset the one-time termination fee recognized in the third quarter. We then expect significant sequential growth in both the first and second quarters of 2027 as contracted deployments enter service. Compute demand is as strong as we've seen, and the range of prospective customers is broad. That gives us choices. Our first priority is delivering the contracts we've already signed. For new deployments, we look for the right combination of customer credit, data center readiness, financing, and returns. We evaluate each deal to earn an acceptable return during the initial contract term. Any value we earn from the equipment after that is additional upside. We use customer prepayments and equipment financing to limit the capital Whitefiber needs to contribute. For signed dedicated compute deployments that have not yet entered service, contracted prepayment covers about 40% of estimated CapEx. The timing of that funding matters as much as the amount. We work to have funds available when equipment prepayments come due. In new customer discussions, were also seeing proposed structures where customers would fund up to 100% of deployment costs upfront. Let me also update you on our previously announced GPU deployment in Paris. Delivery issues with the original equipment manufacturer led us to change suppliers which delayed the deployment. The facility is ready and we received a year in customer prepayments. We're targeting initial service in January and full deployment in February. We do not expect customer penalties related to this delay. Managed services is becoming an important part of our growth strategy. The opportunities we discussed on previous calls are now turning into signed multi-year agreements. We've signed four managed services agreements with a total contract value of about $58 million. Once fully deployed, we expect those agreements to generate about $15 million in annual recurring revenue. Their average initial term is 3.9 years. Based on our current pipeline, we expect managed services to exceed $50 million in annualized recurring revenue by the end of the first quarter of 2027. That depends on signing the additional agreements and bringing those deployments into service. The model is straightforward. Customers fund the hardware and data center capacity. White Fiber deploys, manages, and operates the infrastructure. We don't need to raise debt or equity to fund those assets. We also don't need to take on equipment debt or the risk of what the hardware is worth at the end of the contract. Our agreements can include upfront deployment charges that cover implementation costs and provide a margin. We then earn recurring management and operating fees. These services build on the engineering team, systems, and support capabilities that we already have. Additional deployments will require more resources, but we expect revenue to grow substantially faster than operating costs. That should support very high incremental margins and meaningful cash generation. Our initial agreements are priced at the lower end of the range we're pursuing. As we build a delivery track record, we see opportunities to earn higher fees on new engagements. Those fees will depend on the scope of work and the operating requirements. We're putting more commercial effort into managed services. We also continue to pursue dedicated GPU services selectively where the customer commitments, returns, and financing support attractive economics. In September, we commercially launched White Fiber Continuum. Continuum connects separate data centers so they can function as a single logical GPU cluster. That allows customers to expand beyond the power and space available at one site. We're encouraged by the early response from prospective customers. As power becomes harder to secure at any one location, we believe the ability to combine capacity across sites will become increasingly valuable. I'll now turn the call over to our CFO, Justin, to review the financial results. Justin?

Thanks, Sam. Third quarter revenue was 24 million, up 19% from a year ago, driven by co-location revenue, which more than tripled to 5.6 million. The increase reflected Montreal No. 3, which was not yet bidding a year ago, and the initial contribution from NC1. Club revenue was 18.1 million, including 3.4 million from the previously disclosed customer termination. Excluding that amount, cloud revenue declined approximately 18 year-over-year, reflecting the transition from that customer to new deployments scheduled to enter service over the coming quarters. NC1 contributed $1.1 million of lease revenue during the quarter, beginning when the first 20 megawatt data hole was placed in service on September 23. Earlier billings and installation fees are deferred and recognized over the lease terms Gross profit excluding depreciation and amortization was 13.7 million, a margin of approximately 57% compared with 59% last quarter, mainly reflecting the smaller termination fee. GMA fees were 18 million, down 16% year-over-year. That includes 3.8 million of share-based compensation, lower consulting and transition service fee were partly offset by starter costs that cannot be capitalized. Adjusted EBITDA was approximately break-even compared with positive 2.3 million in prior year quarter. Sequentially, performance reflected the smaller termination fee and the starter costs at deployments not yet generating revenue. Net loss was 101.4 million. They included a 80.8 million induced conversion expense associated with the August repurchase convertible notes, which is excluded from adjusted EBITDA. Turning to liquidity, we ended September with 138.5 million of cash and cash equivalents plus 9 million of restricted cash Deferred revenue, largely customer prepaid, was $306 million, up from $80 million at year end. In August, we issued $310 million of 5% convertible notes due 2032 and used part of the proceeds to retire most of our 2031 notes, netting approximately $180 million. We also retained our rights under the January zero strike call to receive approximately 5.9 million shares subject to the terms. After quarter end, borrowing under the BitDigital facility reached 100 million. We completed the 50 million acquisition of NC2 and NC3 data center and signed a 300 million secure credit facility for NC1. Initial funding is expected later in November or in December. subject to extension, as may be mutually agreed by the borrower and the lenders, and to satisfaction of the remaining funding conditions. Turning to contracted revenue, remaining performance obligations were $729 million as of September 30. During the quarter, we concluded that certain co-location agreements, including those in Montreal 3 and NC1, are classified as leases for accounting purposes. So their future payments are excluded from that figure. And future payment under those leases total approximately $421 million at quarter end. And that figure excludes the second $20 million ANC-1 as that lease has not yet commenced as of September 30.

I'll now turn the call back to Sam. Sam. Thank you, Justin.

Before we take questions, I want to put the NC1 financing in perspective. Signing the $300 million facility is a major milestone for White Fiber. It took longer than we expected, but it puts us in a much stronger position to fund our next phase of growth. Once funded, the facility will allow us to recover part of the capital that we invested in NC1. We can then put that capital to work in our next projects. That is how we intend to scale. Arrange project debt at the outset, invest equity alongside it, deliver the infrastructure, then recycle equity into future deployments. Together with full contracted building at NC1, this gives us a stronger base for growth. In less than a year, We've taken an existing industrial site through a complex retrofit, customer delivery, and institutional financing. That is the model we're building on. And we're bringing that experience to NC2 and 3. We're structuring customer agreements and financing together from the beginning. We have a clear understanding of what lenders require and how to build those requirements into our next leases. The capital we raised in August gives us the flexibility to keep advancing the business through this process. It also reduced our dependence on accessing the markets at any particular moment. Together with the NC1 facility, it puts us in a stronger position to choose our next investments and decide when to make them. The market backdrop has been difficult. AI infrastructure shares have fallen sharply, including ours. We know that has been painful for shareholders and we take it seriously. From where we sit, the opportunity ahead has strengthened. Our signed cloud businesses is larger. Customer interest in our next North Carolina development is broader and stronger. Managed services gives us another way to grow recurring revenue with the potential for very high incremental margins while customers fund the underlying infrastructure. Building and operating this infrastructure is hard. We believe the ability to solve problems and deliver will increasingly distinguish operators in this industry. We now have a stronger platform and a clear path ahead. Our first priority is delivering the business that we've already signed. Strong demand gives us choices about what comes next. will pursue opportunities where customer quality, returns, and financing come together. We're preparing for greater scale while keeping control over the pace of investment. We believe White Fiber's foundation is stronger and its growth prospects are better than at any point in our history. Our job now is to turn that foundation into results, bring contracted capacity into service, turn it into recurring cash flow, Reinvesting capital in attractive new opportunities. That is how we intend to build lasting value for shareholders. With that, we'll open the call for questions. White Fiber President Billy Krassakopoulos and Michael Francisco, Vice President of White Fiber Cloud US, are also joining us for the Q&A. Operator, please open the line.

Thank you. Hi, Nick. Great.

Hi, everyone. Thanks for taking my questions. I first just wanted to ask, just given some of the supply chain issues at NC1, kind of what the learnings are from that and what steps you're taking to de-risk the supply chain as we think about NC2 and NC3, just any color on how much is already on order. That would be great. Thanks.

Billy, the president of White Fiber, who's on the front lines of that, is going to take that question. Billy?

Thanks. Hi, Nick. So, I mean, the NC1 construction build was a very aggressive timeline. Delivering a project at that scale in 10 months and a couple of days was a huge accomplishment for our team, including the supply chain issues. lessons learned. The project timeline and scale for similar projects is going to be a little bit longer, a little bit in advance, pre-ordering the long lead item equipment and strengthening the project management team internally and externally with our contractors and subcontractors. But the playbook that we've been describing for the last few months is the same playbook that we're going to run in NC2 and NC3.

Got it. No, thanks for explaining.

I think it's also worth mentioning there's a lot of synergy with with respect to what we've done at NC1 in connection to NC2 and 3. Using the same workforce and GC and so on is just, we'll be able to apply a lot of lessons learned from NC1 and we could apply that directly just 55 miles away to NC2 and 3.

Absolutely. No, makes sense, guys. Appreciate that.

And then good to see the financing progress at NC1. I guess my question was just how much capital would you intend to recycle from NC1 as kind of equity for NC2 and 3?

I don't know that there's a precise figure we can provide on that. I think you can imagine the majority of the net proceeds from the facility we announced would be earmarked for NCU, the entire NC portfolio expansion. Ultimately, the equity contribution there will depend on the ultimate debt financing and the corresponding and Linda Koss. Speaking of lessons from NC1, I think we were pretty firm in the prepared remarks that we're really endeavoring to structure financing for the next contract well in advance and making sure that it's an investment-grade project and that all the terms are very buttoned up and financeable and able to be underwritten in advance. Ultimately, if we bring on the entire 105 megawatts, say, a little over a billion dollars of a project, then if you pencil in a presumed LTV, presuming the credit markets don't Absolutely deteriorate further. Generally, I think they're better than the headlines. Getting our deal done today, I think it's somewhat of a testament to that. But I think you sort of infer based on the potential LTC what the equity contribution will need to be there.

Okay.

I tend to agree.

Thanks, guys. Appreciate the update.

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

Hi, Paul. Thanks so much for taking the question and congrats on all the progress. I wanted to ask around the cloud business and maybe this one's for Michael. As we think about the commitments across these five contracts and the $750 million in revenue contracted, how should we think about who's bearing the risk across the various sort of constraints in delivery of that capacity as we think about potential reliance on co-location partners other than the deployment you're doing in Montreal. How should we look at that risk profile for delivering the cloud agreements there? And then maybe as a follow on just how the GPU supply chain is looking for the chips themselves, given you, Sam, you noted VR capacity and just trying to get a sense of the backdrop there. Thanks so much.

Thank you. Yeah, so the way that we conduct our business somewhat insulates us from those risks. So in talking about their customers, we've found that they've had challenges in the past with people promising delivery dates, but not at the outset of the deal, putting everything together in a way that's fully transparent to them. The way that we go to market is we do not sign any customer agreements or MSAs until we have clear visibility into the RFS timeline for the data center co-location space, as well as commitments around the GPUs from the selected OEM, as well as the other components necessary to stand up those clusters. So taking that approach sometimes means that we have to disappoint customers at the beginning by pushing our delivery date farther than they might want it in the first place. But it does allow us to actually hit the dates that we are intending to go after. the you know things like Paris happen and but typically for the most part as we go out and sell these deals we only sell things that we are confident that we can deliver in and having a strong set of co-location partners where we have built trust and have confidence in their ability to deliver really helps us meet those timelines.

Great thanks so much Michael and maybe just a quick follow-on on the supply chain there just given how much momentum the cloud business has are you doing any In terms of the build out that you're responsible for on the cloud side, are you doing any proactive long lead purchasing or is this really specific to certain sites with a specific basis of design and so you have to wait until you have more clarity there? How should we think about the long lead, whether it's networking or cooling or any part of that that's your responsibility? Thanks.

Yep, so we have gotten out of doing speculative purchasing this year. So as we adjusted the business model, we've stopped kind of going out and purchasing things with the intent of selling them later. There are occasional exceptions. So in one case, as an example, we did put in orders for NVIDIA networking fabric so that we could make sure that we could deliver NVIDIA RA architectures to our customers. Outside of that, we procure at the point in which we've received the prepayment from the customer, and that helps us kind of determine what our timelines are. Again, going back to my comment about aligning the multiple components, oftentimes there's a piece of that cluster that is the long lead, but we're able to work with our co-location partners to align the timing in the right way. So short answer is we don't do that very much except in unique use cases, and the lead time on the NVIDIA Fabric is so long that we decided to do that knowing full well that that would be sold as soon as we got it.

Great.

Thanks, and congrats again. Thank you.

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

Hey, guys. Thanks for taking my question. Hey, Sam. How are you guys thinking about expected capex for NC2 and NC3? And then specifically, would cost be capped? Where if it comes in a bit higher, you know, there's a percentage that falls to the tenant or some other partner there?

That's probably combined for Billy and Cam.

Do you want to take a first crack either?

Yeah, so we're still in the initial processes of budgeting, and depending on the client selection, the overall project costs could change. And we do expect any customization or additional costs that are required by the end user to be covered by the end user through payment installation fees and prepayments.

Okay, understood. And then my second question is just on the regulatory environment. I believe North Carolina in the elections, the state legislator moved a little bit further left with the state house going towards Democrats. Given the obviously large presence in North Carolina, could you just give us a refresh and kind of the regulatory environment there, whether it's power or permitting?

So go ahead, Billy.

NC2 and NC3 are zoned already for data center use. It's part of our diligence process in selecting these projects, making sure that we can execute them in a timely fashion and minimize all of the risks associated to them.

Again, the project is zoned for data center.

There's no special permits or use authorizations that we need to go out and get for these projects. Community pushback. I mean, to give you an example of what we did in North Carolina One, even though we were under no obligation to do so because we were fully permitted and we were actually in production at the time, we proactively organized community events as part of our commitment to building positive relationships with the community. That event had just over 200 people that attended. And we explained our retrofit strategy of repurposing older warehouse, our closed loop water system, which, by the way, was part of our standard practice before water use was even a talking point in the media. And we also discussed our contribution to the community and all that was greatly appreciated. We plan to do the same thing in NC2 and in NC3.

Okay, understood. Thank you for that, gentlemen. Appreciate it.

We will take our next question from Michael Donovan with CompassPoint.

Hi, Michael. Hi, Sam. Hi, guys. Thanks for taking my question.

Last quarter, you mentioned late-stage customer opportunities tied to Cranboo. How have those Discussions progressed, and when do you expect the first capacity to be available for deployment?

So those discussions have progressed very nicely. We expect the first deals to close in early Q1, maybe late Q4, with deployments happening in Q1.

Appreciate that. And then on continuum, since

Launching Continuum, how would you characterize customer interest and what types of customers are most engaged?

So Continuum is an interesting product because it is relevant to a fairly niche audience outside of including folks such as ourselves, but also that there are some enterprise customer interests that we've seen as well. Most of the folks that we're seeing are folks that are looking to solve for the distributed or fragmented power problem. And so we've been engaged with colocation providers that are focused on kind of regional data center setups who are interested in the technology to be able to bring those kind of regional hub and spoke solutions to bear, as well as some enterprise customers that have looked at this in terms of being able to provide a more robust kind of logical supercluster leveraging Again, smaller sets of power. I anticipate that the first deal we close on this is going to take a little bit just due to what it actually takes to deploy and plan and kind of make it a reality. But we have seen good initial interest and have engaged in some very interesting conversations about scaling this over time.

That's helpful. Ed, are you thinking about using that to link NC1, NC2, and NC3?

Or would this be for other internal sites?

It's a consideration that we could make depending on customer interest, but I can't comment on the structure of our current deals or current negotiations whether or not that would be included in that.

I appreciate that. Moving forward. Thank you.

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

Hi, George. Thank you. Sam, you mentioned that your demand for NC2 and 3 is substantially stronger than what you were seeing for 1. Can you just give us a sense, is that a site-specific situation, a market dynamic? Is that a comfort in your ability to deliver with a retrofit model? Just curious on that.

I think it's the first and third, but I'll let Billy answer that question.

Yeah, so multiple factors here. What we did in NC1 resonated across multiple people that were looking at the project previously. The ability to deliver power in 2027 is a major factor. Many companies are delaying or canceling projects, and we've shown that we can deliver in an extremely quick turnaround and looking to replicate that across our three projects in North Carolina in 2027. So potential customers appreciate that and they see the track record that's been proven and that's what's driving the demand at those sites.

I understand. Sam, you gave a pretty compelling statistic that in some cases you might see 100% upfront payments I just figured it couldn't go without being asked on this call. What you meant by that?

Michael, you want to take that?

Sure. So as we see the markets tightening or at least see the headlines around the markets tightening, I tend to agree with Cam that the headlines paint a more dire picture than what the reality is. There's a number of factors that come into play. There's both the factor that a number of customers are looking at owning that hardware on their balance sheet as an advantage. So thinking about it in the context of traditional enterprise hardware acquisition, where they're having it on the balance sheet actually helps them from a taxes and depreciation standpoint. There's some accounting benefit there. But at the same time, and we're seeing that with enterprises as well as some of the larger foundational labs, some of the larger well-funded AI startups are also, we're seeing a trend on that front as well, where they're looking to take some of the capital that they've built up to invest in their own hardware to secure that over time and make sure that they have those assets readily available. At the same time, I think we are also seeing customers understand the reality of the market and understanding that by coming in and helping with the prepayment, it actually facilitates the transaction more rapidly, gets them up and online more quickly, and then kind of reduces the cost of the OPEX across the term of the deal. So there's a number of factors that are contributing to that, and as the market continues to mature and adjust, I think we'll see both more of this, but I think people are also going to figure out some of the financing bits as well to get creative with how we deploy these things.

Very insightful. Thank you for the thoughts. You're welcome.

We will take our next question from Mike Colanese with HC Wainwright.

Hi, Mike. Hey, Sam and Tim, good morning. Thank you for taking my questions. First one for me, I was just curious if you could provide more color as to what the remaining milestones are outstanding that need to be achieved before you can convert existing LOIs at NC2 and 3 into signed leases now that the acquisition is closed.

Billy, do you want to take that?

Sure.

revisions of the base of design for the customers that we are currently speaking with, validation of the delivery times for the long lead item equipment are the two main driving points, milestones that are currently being discussed with potential clients to execute on a service agreement.

That's helpful, Chloe. Appreciate that. And then it was discussed a bit earlier around the aggressive deployment timeline. You guys were able to execute just over 10 months on NC1. I'm curious, based on the basis of design under negotiation, what could the deployment timelines look like for NC2 and 3 here, especially given that, you know, the initial RFS is targeted for 2027, so just trying to better understand when you could actually monetize potential contracts once a lease is executed.

Yeah, so right now we're looking at end of 2027. giving ourselves a little bit more buffer. Again, already aggressive timeline. I don't see many companies executing these types of deployments in less than 12 months. We're targeting the same playbook, even though the base of design is a newer generation of chips or a different manufacturer altogether, depending on who the end user will be. We still are confident that we can deliver a project in 12 months.

And Mike, can I just add, we're just not really overly eager to optimize for time. I think we're trying a pretty firm line in the sand that we're optimizing for holistic project returns and the ability to very easily and cleanly finance a project. So we're just, we're really not eager to compromise on any, you know, Thank you.

We will take our next question from Ben Summers with BTIG.

Hey, good morning, and thank you guys for taking my question. So you made a comment on enterprises and government and whatnot becoming, you know, a bigger part of the demand mix moving forward potentially for co-location services. Just kind of curious how that dynamic is potentially playing a role and how you think about expanding the power portfolio from here.

Is that for cloud or co-location?

This is in reference to co-location?

Yeah, but for both. I mean, for either or.

Yeah, just depending on who you're speaking to. So with respect to co-location, we are looking at Fortune 100 companies and even public sector end clients. Is there any more that we can talk about, Billy, without giving away our current negotiations?

Probably not.

A little difficult, but I mean, the scarcity of capacity, again, is a driving factor there. The larger These larger portions of capacity are being taken by the hyperscalers and smaller colocation megawatts. These clients are kind of left in the water without any available inventory in the market. So we have gotten a handful of potential

clients that are in the top Fortune 100 that are looking for smaller deployments of 10, 20 megawatts.

Super helpful. And then on the cloud services side, you guys noted, I want to say an average contract length of around four years. Just kind of how do you think about So our primary guidance here is that we do not structure contracts that do not have a full payback within the term of the original contract itself.

What we're seeing from customers is actually an interest in longer term contracts than what we saw last year or even earlier this year. And I think there's a couple of dynamics there. One of the key dynamics is simply that I think it used to be believed that you could go negotiate a two-year contract for GPUs and then have an out after two years to renegotiate and get those prices down. I think what we've seen in the market is that's not actually the reality, that sometimes if you look at the cost of the H100s now, the cost of those GPUs is actually higher today than it was before. when they launched. And if I had negotiated a contract for two years ago, I'd be paying more at the point of renegotiation. So those dynamics aren't playing out for those folks that were looking for shorter-term deals. The other piece of this is that it's difficult to find GPU capacity. And so the idea that you want to sign a shorter-term contract to go back through the process to maybe pay more or maybe not even be able to find that capacity three years from now is not a risk that a lot of folks are willing to take. So we're seeing them push for longer terms. And with those longer terms, we're able to be a little bit more creative with the structure of the contract to make sure that we get our payback, but then provide some relief in terms of the average cost for the customer across the term of that deal by kind of blending the rates across that lifecycle.

Super helpful. Thank you for taking my questions.

You're welcome.

We will take our next question from Brian Baratian with Roth Capital Partners.

Hey, guys. Thanks for taking my question. Hey, Sam. So, great, like, tremendous progress on the corridor. Thank you for a really informative print. My main question when you're talking about forward leases, which I fully agree with, by the way, is, you know, targeting more of an investment grade or an investment grade, you know, wrapper. As we think about your future capacity, Maybe you could talk a little bit just so I understand. Is that like the same format that you have at NC1? Or do you mean for the wrapper situation? Like, is that separate for what we'd expect to see at NC2 and NC3? Because obviously, these are triple net lease real estate assets. So your cap rate, your underwritten cap rate is obviously going to be lower for a higher IG offtake. So could you give a little color on that?

I'll defer to Billy and Cam.

I'm not sure I followed the question 100%, but perhaps Billy and Cam did.

Yeah, I can chime in. Hey, Brian. It's Cam. Yeah, I mean, holistically, I think we're just taking a much different approach than we took for our initial NC1 lease, where, like, on a prevailing basis, there's not a full investment credit wrap. It's not directly with an IG counterparty. and as you can see just it was a longer process of putting in debt capital and that's clearly something we're looking to solve so one of our main goals something we're not really willing to compromise on is ensuring the next lease we sign is rated investment grade and that we have you know a certain level of comfort pre-lease signing that there is and some sort of committed debt at attractive terms.

Great. Thanks, Cam. And then the other thing is, you know, in the face of rising costs that we're seeing in supply chain that we're just seeing across the industry, a tougher AI infrastructure backdrop, you know, you mentioned that being able to, you know, pass on costs and achieve like higher than like higher yields or that spread to your cost of capital to your yield on cost. have customer conversations or preliminary customer conversations, have they been receptive to that, saying, hey, like, cost of capital is going up, our build costs are going up, the demand for power is so strong that they're saying, hey, you know, the dollar per kilowatt month that you guys are charging, it's fine. We can pay an extra 5, 10, 15, 20 dollars. of Rent. Has that been received well by customers?

Just before we answer that question, I just want to say that it's going to be – we're in current – negotiations are ongoing right now, and we just don't want to disclose, like, the current postures of various counterparties. So, I would speak in generalities, either Cam or – That's what I – Yeah, that's specifically what I mean.

Right.

Yeah, generally, we don't see any problems there. Customers are actually offering better prepayments and installation fees, again, as a sort of premium to delivering capacity in 2027 because of where we are at in the industry right now on available capacity and inventory. And everyone understands that the costs of labor, equipment, Electricity, everything, all these costs have increased since, have outpaced the increases, normal increases in the last year, and it hasn't been a problem in negotiations. Excellent. Congrats.

I think the sweeping generalization picture, but the comparison that we've made is that year over year, as we evaluate customer proposals and our own budget at this time for the development. The implied yields are more attractive than what we were seeing when we signed our initial North Carolina lease.

Perfect.

Outstanding quarter and great progress, guys. Looking forward to the next quarter. We are too.

We will take our next question from Kevin Deedy with Jones Trading.

Hi, Kevin. Hey, Sam.

Thanks for having me on, and great to see you in Singapore, too, so thanks for being there. I'm curious about MTL2. Congrats on getting that one going, I guess, but it It's unclear what power he had available there. I've forgotten. Please excuse me. You also mentioned, I think, something about behind the meter, and I'm just kind of wondering how that process would work given pretty tight regulation strings in Kibbutz.

Is that you or the dog asking the question?

Yeah, that's the dog. That's not me. but I guess that's a good question, depends on who you ask.

Right. Billy, do you want to take that or Kim?

Yeah, so that was a smaller deployment from the get-go. There's just about four megawatts available at the site and we are evaluating natural gas fuel cells to help us produce a little bit more energy and a little bit more heat that we can recover and and some heat reject projects that we're looking at in the area that would put us in a better spotlight to get higher allocations from the utility company as well. So it's a kind of a two in one solution for us in implementing natural gas fuel cells at that facility.

Okay. Thanks, Billy.

On the full 99 megawatts at NC1, is there any change in the timeline when that could possibly be delivered? And maybe add some color in your negotiations with Duke in taking NC2 and 3 from 60 to 200.

The NC2 and NC3 is still in preliminary talks with engineering on the expansion capabilities. Like usual in our due diligence, we make sure that there's a positive way forward to getting more power. So those are still in initial talks with Duke. And for NC1, same thing, we're still looking at some supply chain issues on the Duke side where we're working together, where we can help and hopefully advancing that timeline. But right now it's looking at the end of 2027.

This might be a sore topic, but I'll leave it to you, Sam, to add whatever you think is appropriate. Obviously, there were some struggles with Microsoft dealing with N scale, and I'm just wondering if any of that may have rippled through to you and the prospect of N scale filling out all the available power as it comes to NC1.

All right, go ahead, Billy, but that's not, that's not to work. Go ahead.

I mean, it's hard to comment on a relationship that we don't have any Any insight on what we know is that Nscale is using the capacity at NC1. The expansion capacity, I mean, although we would for continuity and for client relationship, we would like for it to be Nscale for the expansion. But we're open to other users and as Cam said, Getting IG-type terms is the main driving factor in our selection process. But what we can say, what we do know, is that N-scale is using the current capacity at NC1.

Perfect. Thanks, Billy. Thanks for taking the questions, gentlemen.

Thank you.

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 everybody for joining us today. We look forward to the next earnings call and have a wonderful day. Thank you.

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