Logo of jester cap with thought bubble.

Image source: The Motley Fool.

DATE

Monday, Aug. 10, 2026 at 8:00 a.m. ET

CALL PARTICIPANTS

  • Investor Relations - Laine Yonker
  • Director and Chief Executive Officer - Ben Gagnon
  • Chief Financial Officer - Jonathan Mir

TAKEAWAYS

  • Total Liquidity -- $819 million as of Aug. 7, 2026, comprising approximately $698 million in unrestricted cash and $121 million in unencumbered Bitcoin.
  • Revenue -- $30 million, representing a 50% decrease compared to $61 million reflecting lower average Bitcoin prices and the shutdown of Moses Lake mining operations.
  • Operating Loss -- $141 million, compared to $11 million in operating income in the prior year period due to $63 million in accelerated depreciation from mining rig shutdowns.
  • Non-GAAP Adjusted EBITDA -- negative $24 million, down from $7 million driven by higher general and administrative expenses and increased stock-based compensation.
  • Convertible Senior Notes -- $458 million raised to fund power capacity expansion at existing sites including Panther Creek and Scrubgrass.
  • Bitcoin Liquidation -- $75 million in proceeds from selling 1,085 Bitcoin between April 1, 2026, and Aug. 7, 2026, as part of a planned exit from cryptocurrency.
  • Remaining Bitcoin Holdings -- 1,861 Bitcoin as of Aug. 7, 2026, with management intending to liquidate the entire position by the end of 2026.
  • Cash SG&A Guidance -- $100 million for the full year 2026, reflecting the hiring of senior subject matter experts to support the commercialization phase.
  • Panther Creek Capacity -- 350 megawatts of secured utility capacity with potential expansion to 500 megawatts or more for high-density deployments.
  • Panther Creek Timeline -- ready-for-service date remains 2027 despite a multi-month delay in final environmental permitting processes.
  • Sharon Capacity -- 110 megawatts to be delivered in a single phase, reflecting a consolidated design for simpler construction and higher power density.
  • Sharon Permitting -- land development approval secured during the quarter with only a few environmental permits remaining.
  • Moses Lake Timeline -- ready-for-service expected in 2027, representing a delay of approximately two months from original guidance.
  • Moses Lake Scope -- 18 megawatts total capacity, as the company opted to relinquish a 10-megawatt expansion option to focus on existing infrastructure.
  • Scrubgrass Load Study -- 750 megawatts for a potential 1 gigawatt plus campus, currently in the energy application stage with utility partners.
  • Sherbrooke Expansion -- 96 megawatts of high-performance computing power capacity approved by local utilities, pending final provincial sign-off.
  • U.S. Bitcoin Operations -- 100% decommissioned during the second quarter to prepare sites for transition to AI and high-performance computing infrastructure.
  • Capital Position -- liquidity supports site development through 2028, covering lease signing and expansion capacity opportunities.
  • Supply Chain Execution -- accepted delivery of long-lead items, including transformers at Sharon and Vertiv modules at Moses Lake.
  • Net Loss -- $65 million, including a $64 million loss from continuing operations compared to $13 million in income in the prior year period.

Need a quote from a Motley Fool analyst? Email [email protected]

RISKS

  • Gagnon stated, "the final process with regulators is taking a few months longer than originally anticipated," regarding environmental permits at the Panther Creek site.
  • Gagnon noted, "Moses Lake has been delayed maybe a couple of months since our original guidance," while indicating it remains the first site expected to come online.

SUMMARY

Management reported a strategic transition toward becoming a dedicated high-performance computing and AI infrastructure provider. The company stated that power availability remains the defining constraint for technology development, leading to a commercial focus on securing energy capacity in key North American markets. Management indicated that permitting is advancing across all primary sites, with multiple potential tenants engaged in negotiations for near-term power. The company noted that its current liquidity position is intended to support project development through lease execution and expansion efforts.

  • CEO Gagnon reported that Moses Lake will be the first site to come online and is expected to "return significant equity capital to our balance sheet and become our first source of durable free cash flow."
  • The company appointed Ganesh Aiyer as President to lead commercial efforts, citing his seven years of experience as Chief Business Officer of Digital Realty.
  • Management noted that power scarcity has shifted the commercial process from a sales pitch to a negotiation, with Gagnon stating, "holding the bottleneck everyone needs to grow means we are negotiating from strength."
  • The company transitioned its internal operations to Claude for enterprise AI, with Gagnon noting the importance of remaining model-agnostic because "the market is going to continue to change and adjust."
  • Management characterized Pennsylvania as a unique "centrist state" with a second-mover advantage in data center development due to its existing industrial and energy framework.
  • In Quebec, the company intends to consolidate legacy Bitcoin power agreements into a single 96-megawatt high-performance computing agreement to capitalize on captive demand and data sovereignty requirements.

INDUSTRY GLOSSARY

  • AI: Artificial Intelligence.
  • BESS: Battery Energy Storage System.
  • CCGT: Combined Cycle Gas Turbine.
  • DEP: Department of Environmental Protection.
  • ESA: Energy Service Agreement.
  • HPC: High-performance computing.
  • IPP: Independent Power Producer.
  • PJM: A regional transmission organization that coordinates the movement of wholesale electricity in all or parts of 13 states and the District of Columbia.
  • PPL: PPL Corporation, an energy company headquartered in Allentown, Pennsylvania.
  • RFS: Ready-for-Service.
  • Vertiv Modules: Pre-fabricated data center power and cooling infrastructure components.

Full Conference Call Transcript

Operator: Ladies and gentlemen, thank you for standing by. Welcome to Keel Infrastructure Corp. Second Quarter 2026 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to turn the conference over to Laine Yonker, Keel Infrastructure Investor Relations. Please go ahead.

Laine Yonker: Thank you, and welcome to Keel Infrastructure's Second Quarter 2026 Conference Call. With me on the call today are Director and Chief Executive Officer, Ben Gagnon; and Chief Financial Officer, Jonathan Mir. Before we begin, please note, this call is being webcast with an accompanying slide presentation. Today's press release and presentation can be accessed on our website under the Investors section. Turning to Slide 2. I'd like to remind everyone that certain forward-looking statements will be made during this call and that future results could differ from those implied in this statement. The forward-looking information is based on certain assumptions and is subject to risks and uncertainties.

I invite you to consult Keel's 10-Q for a complete list, which will be available on our website and the SEC website. Please note that references will be made to certain non-GAAP financial measures, and therefore, may not be comparable to similar measures presented by other companies. We invite listeners to refer to today's press release and our filed 10-Q for definitions of the non-GAAP measures and their reconciliations to GAAP measures. Please note that all financial references are denominated in U.S. dollars, unless otherwise noted. And now turning to Slide 3. It is my pleasure to turn the call over to Ben Gagnon, member of the Keel Board of Directors and our Chief Executive Officer. Ben, please go ahead.

Benjamin Gagnon: Thank you, Laine, and good morning, everyone. 18 months ago, we laid out a clear vision for both Keel and the data center industry. We told you that the defining constraint of the most important technology of our lifetime was not chips or capital, it was power. And we told you that by the end of 2026, power would be even more constrained and even higher demand. We laid out a clear investment thesis that focusing on developing power in the right places on time lines that matter would be incredibly valuable to prospective tenants and value maximizing for shareholders.

We explained the necessary work ahead of time, and we kept you informed step by step exactly how we would transform this company into a premier regional data center developer. We said we would exit Latin America and Bitcoin and become an American HPC and AI company. We did. We said we would rebuild the balance sheet to enable our transition to an HPC and AI infrastructure company. We did. We said we would be ready to monetize our assets when power was scarcer and demand was stronger. We are. Throughout this transition, we've delivered on our commitments, either on time or early.

If time lines moved, we told you why, we told you what it meant, and we told you what did and did not change. That's not luck. That's a track record reflecting strategic discipline and consistent execution. Turning to Slide 4. In May, we shared that management was focused on 3 things this year: one, advancing permitting and leasing across all 3 priority sites; two, securing our expansion capacity; and three, delivering energized megawatts as quickly as possible for our customers. 90 days later here is where each one stands. First, on permitting and leasing.

I will walk through each site's permitting and leasing update individually in a moment, but I'd like to first highlight the main takeaways here: one, we further advanced permitting across all 3 priority sites this quarter and have clear visibility on permit completion at each site; and two, near-term power is scarce and our sites have it. That scarcity is doing the work for us. It's why all 3 sites have multiple potential customers engaged and negotiating, and it's why these conversations start from a very different place than they would have 2 years ago.

This is an important distinction because when your sites solve the hardest and most valuable problem potential tenants have, power, timing and location, the commercial process stops being a pitch and it starts being a negotiation. And so to lead this next pivotal phase, last month, we welcomed Ganesh Aiyer as President of Keel. Ganesh has spent his career at the intersection of infrastructure and commercial strategy and joins us after nearly 7 years as Chief Business Officer of Digital Realty. He is now leading our commercial efforts. And while he has only been with us about a month, he has already hit the ground running. Second, on expansion capacity.

Last quarter, we explained our thesis that the market was not ascribing much value to the unsecured megawatts in our expansion capacity. We also explained that securing these megawatts was an important focus for management and a key value driver for shareholders. So first, in Pennsylvania, we've been working closely with both of our utility partners to advance our power applications for expanded capacity. While we can't provide details today, we are increasingly confident in our ability to convert potential expansion capacity from our 2-gigawatt Pennsylvania pipeline into more signed ESAs, delivering energized megawatts for HPC through 2030. We expect we should be able to provide investors with a fulsome update as early as December or January.

Additionally, we advanced our Sherbrooke data center plans during the quarter, securing all necessary local approvals from the city and the local utility, with only provincial approval outstanding. We are excited to significantly expand our relationship with Sherbrooke, where, over the past 7 years, we have generated substantial revenues, taxes, jobs and community benefits. If approved, we will consolidate our 3 legacy Bitcoin power purchase agreements into a single 96-megawatt HPC and AI power purchase agreement for a new data center development in Sherbrooke, a market where new data center energy capacity is nearly impossible to secure and is in high demand.

Sherbrooke will be designed from the ground up to support the next generation of hardware and has the potential to become one of the most technologically advanced data centers in all of Quebec upon completion. Third, delivering energized infrastructure as soon as possible. Every commercial negotiation comes down to the same 2 questions: how fast can I get my first megawatts, and how fast and far can I keep growing with you after that? So in parallel with every commercial negotiation, we are working constantly with our partners, our manufacturers and our supply chains to protect the time lines our customers are underwriting. Turning to Slide 5.

Let me share some examples because most of this work never makes a press release. This quarter alone, we accepted delivery of long lead time items and the first Vertiv modules at Moses Lake. And we'll be conducting further factory and predelivery inspections with Vertiv as modules come off the assembly line. We completed inspections for backup generation equipment in Moses Lake. We took delivery of several long lead time items in Sharon, including multiple transformers. We began executing final fiber contracts across our 3 sites, ensuring multiple path redundancy and connectivity will be available before the sites are online. We continue to update our data center designs, improving power density specs so that we can meet customers' hardware requirements.

We completed the first phase of construction across all 3 sites, which is the decommissioning of all U.S. Bitcoin mining operations. And most importantly, we significantly deepened our bench of subject matter experts across construction, power, fiber, engineering, controls and other critical disciplines, and we continue to add talent in these areas. Clear deliberate steps to derisk our project time lines and ensure we can deliver state-of-the-art infrastructure within the time frames and budgets our customers require. These steps mark the difference between a promise and a delivery date. Step back and look at what all of this adds up to. 18 months ago, we laid out our thesis and our strategy.

Today, we are exactly where we said we wanted to be. The market is where we anticipated it would be. We are now active in the commercial process with the sites we wanted to bring to market, at the moment we wanted to bring them to market. We are doing so from a position of financial strength and with permitting largely derisked. We followed through on our promise not to cap upside by signing leases prematurely and that patience is now paying for itself. This is our goldilocks phase, not too early to matter, not too late to win, exactly the window we built this company to hit.

Now let me show you what execution looks like on the ground starting at Moses Lake. Turning to Slide 6. Moses Lake is shaping up to be a milestone site for Keel. It will likely be the first sight fully permitted, the first site to come online, the first site to generate HPC revenues, and upon commissioning, we expect it to return significant equity capital to our balance sheet and become our first source of durable free cash flow. Permitting in Washington works a bit differently than in Pennsylvania, and has allowed us to start site development while we finish the go vertical permitting process, which we expect will wrap up later this quarter.

The Bitcoin mine that stood there before is gone, completely removed. Today, the site is being prepared for the Vertiv modules with every piece of critical long-lead equipment secured and being actively manufactured. In fact, the first Vertiv modules have already arrived on site with deliveries continuing from here. When you look at that rendering on the slide, understand that everything in it is bought, contracted or already being manufactured, including the building itself. We look forward to delivering Moses Lake as our first fully commissioned and energized data center in 2027. And the commercial process reflects this.

Moses Lake has interest from exactly the potential tenants you would want, leading AI companies, GPU clouds and enterprises that need power now. Inbound activity and negotiations have accelerated throughout the quarter, reflecting just how scarce near-term power is in the Pacific Northwest. Moses Lake serves a different customer profile than our Pennsylvania sites, faster-moving companies that value speed and a fully operated facility. So due to that customer demand, we may structure leases here on a modified gross basis rather than triple net with credit support structured to match.

That approach lets these tenants move at the speed they need, keeps Keel on operational control and creates more value for a site with the size and scope of Moses Lake. Turning to Slide 7. As Sharon momentum continues to build, we secured a full zoning in April. Land development was approved during the quarter and our final environmental permits are submitted and progressing on track with only a few environmental permits remaining before Sharon is cleared. We also iterated on the designs throughout the quarter, evaluating how to best consolidate the compute capacity, which we believe would be a simpler, less complex build and an overall stronger product.

Sharon is in active commercial discussions today with multiple parties engaging on the site simultaneously and evaluating it for exactly what it is, rare, uncontracted 2027 power in PJM. The structures under discussion here are focused on triple net and include pairing fast-growing AI companies with investment-grade credit support, exactly the kind of structure that enables a high-growth customer to deliver a financeable long-term lease. Turning to Slide 8. And then there's Panther Creek. 350 megawatts of secured utility capacity with PPL, 2 hours from New York and Philadelphia in the middle of one of the most sought after AI corridors in America.

This quarter, we secured zoning, we secured conditional land development approval and we refined the data center design for higher density deployments because with potential expansion capacity to 500 megawatts or more, that is where customer demand is going, not just solving for near-term power, but power that can keep scaling for years to come. On permits, we are in the final stages of our last few environmental permits. All have been submitted and are progressing. However, the final process with regulators is taking a few months longer than originally anticipated.

For investors, I would like to clarify what this means: one, the final DEP permitting does not change our planned power delivery schedule under the ESA; two, it does not change the anticipated economics of the project; and most importantly, three, it has not slowed commercial progress or interest. As of today, our earliest RFS date continues to be 2027. And for the customers that we are speaking to, we don't believe this will have an impact. Commercial interest at Panther Creek is high, and we believe recent broader market dynamics are also beneficial for the site.

Because of the scale of the Panther Creek campus, engagement is led by large, sophisticated AI companies, and we expect interest from the very largest players to deepen as the site reaches execution-ready status on permitting. That is the pattern in this market. The bigger the counterparty, the more they value certainty. And with every permit that lands, Panther Creek becomes something only a handful of sites in America can offer, near-term power, at scale with room to keep growing for years. Today, we have multiple potential customers negotiating across multiple sites simultaneously. Interest across the portfolio far exceeds the capacity we have to lease.

And these are the counterparties you would want at the table, hyperscalers, leading AI companies, GPU cloud and large enterprise. While I cannot name names or reveal particulars, I want you to understand that there's competitive tension in this process and our challenge is not finding customer demand, but in choosing among it. I also want to be direct about how we think about signing. We have been very clear for the past 18 months about our commercial time line. We did not rush to the finish line, but rather took the time to derisk our sites, build commercial interest and ensure we secure the best economics possible for our shareholders.

A lease is not a trophy for a press release, it is a 15-year commitment of infrastructure, credit and trust. And the difference between a good lease and a great one is measured in hundreds of millions of dollars over its life. Holding the bottleneck everyone needs to grow means we are negotiating from strength, and we will focus on optimizing across customers, economics and cost of capital. We are not going to cap the upside of a generational asset in order to deliver a headline. We remain very optimistic and increasingly confident from the engaged and active tenants in our commercial process.

The intensity makes clear that our portfolio is exceptionally well positioned to solve a wide variety of customers' problems. Secured power available in 2027, attractive locations and proven delivery partners remain the differentiators driving every customer conversation we're having. Turning to Slide 9. And with that, I'll turn it over to Jonathan to discuss our Q2 financial results.

Jonathan Mir: Thanks, Ben, and good morning, everyone. I'd like to open with a simple message reiterating what I communicated on our Q1 call. We are better capitalized today than at any point in this company's history and that capital position gives us something invaluable in this market, the ability to both advance and derisk our sites at the pace our customers require and to make commercial decisions driven by our objective of delivering the best possible long-term risk-adjusted shareholder returns rather than being driven by time-pressured liquidity position. I'll walk through our capital strategy in more detail, but first, I'll review our Q2 results. Turning to Slide 10.

For the second quarter of 2026, revenue was $30 million compared to $61 million in the second quarter of 2025. The change was largely due to the decrease in average Bitcoin price and the shutdown of the Moses Lake cryptocurrency mining operations during the quarter. Operating loss for the quarter was $141 million compared to operating income of $11 million in the prior year period. This change includes $63 million of accelerated depreciation relating to mining rig shutdown at the Panther Creek and Scrubgrass sites, change in fair value of Bitcoin and realized loss in Bitcoin was $20 million compared to a gain of $32 million in Q2 2025.

Loss from continuing operations of $64 million or $0.11 a share compared to income from continuing operations of $13 million in Q2 '25. Adjusted EBITDA for the quarter was negative $24 million compared to $7 million in the prior year period. This decrease in operating margins reflects a decline in the Bitcoin price and increase in G&A related to senior subject matter expert hires as we scale up to the next stage of our business and an increase in stock-based compensation year-over-year. Our cash SG&A for the first half of 2026 and averaged $23 million per quarter, and we are currently tracking $100 million of cash SG&A for the year.

Again, the increase versus prior year is driven largely by the high-quality selective senior hires needed to support the commercialization phase of our strategy. The company sold 1,085 Bitcoin for $75 million in proceeds during the period beginning April 1, 2026, and ending August 7, 2026. As of August 7, 2026, the company's Bitcoin balance stands at 1,861 Bitcoin. As previously discussed, our intent is to liquidate our Bitcoin position in 2026. Turning to Slide 11. I'll now cover some capital market observations as well as the liquidity update. In June, we closed a $458 million offering of convertible senior notes upsized from an initial $350 million, having received strong investor demand, which we greatly appreciate.

This investor demand allowed us to be thoughtful about who we brought on to our cap table, and we're pleased to have added several high-quality, long-term oriented investors as a result. Investor feedback has been positive regarding our clarity on how we will use this new capital. This isn't discretionary or speculative capital that is earmarked to expand power capacity at 2 of our derisked owned sites, Panther Creek and Scrubgrass. We're not using these proceeds to take on new development risk. We're using them to build incremental power capacity at existing sites.

Whenever we need external capital, our commitment is to be clear on the uses of that capital and why we believe the associated long-term risk-adjusted returns create value for our shareholders. Taken together, we see the convert offering as having been both a vote of confidence from the market and the direct enabler of the next phase of our strategy execution, including pipeline growth through expansion capacity. Moving on to liquidity. Total liquidity as of August 7 was $819 million compared to $533 million reported at the beginning of May. To reiterate, we believe our current liquidity supports site development through lease signing, expansion capacity opportunity and fully funds our cash SG&A through 2028.

Before we open the call to Q&A, let me touch on observations about capital markets conditions as they bear directly on how we plan to fund construction of our sites. First, in respect to the project level high-yield debt financing, we're comfortable with current market conditions. Even with spreads widening, we believe there is adequate depth for the amounts we would raise and prospective returns to equity capital remain attractive. Second, an investment-grade offtake directly or wrap remains critical to obtaining efficient debt financing. The cost of financing against a noninvestment-grade partner is meaningfully higher and has less market depth.

However, at least for now, capacity is available in the market to finance both against investment grade and selective noninvestment-grade customers. We continue to believe that an investment-grade customer wrap with durable lease terms is the best choice for shareholders in those circumstances. Lastly, our liquidity position enabled us to evaluate any potential capital requirements on a post-lease basis when we expect our cost of capital to decrease. In summary, we believe that current market conditions leave us well positioned to finance each site's construction smoothly and on terms that will create value for our shareholders. Turning to Slide 12, I'll turn it back to Ben for some closing comments.

Benjamin Gagnon: Thank you, Jonathan. Before we open the line for questions, I want to say a quick word about why Keel is doing all of this. Every generation builds its defining infrastructure, and it always gets built before the world agrees it should be. The railroads, the electric grid, the highways, the Internet, intelligence is ours. Work is no longer measured in jewels, it is measured in tokens. And while the price of a token has a ceiling, the value of one does not. We named this company Keel for a reason. The infrastructure we are building is the foundation that enables the next generation.

We are not competing with anyone's ideas about AI, we are powering the people who have them. 18 months ago, this was just a thesis for Keel. Today, we are a company executing in exactly the window we saw coming. Operator, please open the line for questions.

Operator: [Operator Instructions] And first question is going to come from Gareth Gacetta with Cantor.

Gareth Gacetta: it's Gareth on for Brett. I was hoping you could touch on kind of the political environment around data centers kind of across the U.S. I know you mentioned that these kind of developments haven't really changed the power delivery schedule or also kind of the commercial progress among potential tenants. But can you just talk about how these potential tenants are looking at the regulatory backdrop and what that might be impacted on their timeframe?

Benjamin Gagnon: Yes, happy to do that, and thanks for the question, Gareth. The regulatory backdrop and the political backdrop is something that we obviously are watching very, very closely. Clearly, there's a lot of headlines around the U.S. right now with moratoriums and regulatory actions and kind of new frameworks or new policies or new tariffs that are being proposed and being suggested. I think the reality is, is that every time that, that happens in a place, it's going to increase the value of the other sites that are not impacted by those regulations. And obviously, in a market dynamic where there's so much growth happening so fast, sometimes some markets need a little bit of time to catch up.

I think one of the advantages that we have here in Pennsylvania is Pennsylvania is kind of enjoying the second mover advantage. It definitely wasn't the first to jump up there and start building data centers. They've really had a lot more permits and rules and different steps and hoops to jump through in the first place. And so I think that the reactions that you're seeing across the country are due to the huge influx of data center demand in applications. And I think Pennsylvania had a pretty good framework in place already for large industry, large manufacturers, very large kind of consumers coming in to build industrial capacity.

And I think it sets us up, and I think it probably can create some value to Pennsylvania to see these actions taking place in other sites because that capacity still needs to come in the United States. And those are the areas that there's going to be continued opportunity in.

Gareth Gacetta: Great. That's super helpful. And then maybe just a quick follow-up. Could you touch on your current pipeline? I think it's about 480 megawatts you guys have secured. But could you just provide any color on how much of that pipeline is exposed to this application process?

Benjamin Gagnon: So we've got 2 different buckets of energy. We've got our secured and we've got our expansion capacity. As of right now, all of our secured capacity, we believe is unimpacted to date, and we're going to continue to monitor that very, very closely. The expansion capacity may be impacted by future changes or future policy implications. But right now, everything is progressing incredibly well on securing our expansion capacity. I mentioned it briefly on the call. We're working with our utilities on a daily basis. Our applications to secure our expansion capacity, which is almost 2 gigawatts across the state, is going very, very well.

And we're increasingly confident that we're going to be able to secure additional power and look forward to giving investors the update as early as December or January.

Operator: And our next question will come from Greg Lewis with BTIG.

Gregory Lewis: I was hoping to kind of talk a little bit about the permitting process. I noticed you talked about some of the environmental permitting, just I's that we have to dot and Q's we have to cross, as you're working with your data center [ customers ], I'm curious, is there like a dual process around how we could address some permitting issues? And the reason I'm asking is one of the things that we've heard is sometimes the backup power generation, if it's diesel or natural gas, tends to trigger some environmental permitting challenges or just things we need to address versus maybe using backup batteries as a solution.

Just kind of curious if that's something that we're exploring just in case the environmental permitting takes longer or is just a slower moving process maybe than we thought?

Benjamin Gagnon: Yes. Thanks, Greg. So to answer your question, you're certainly right that when going for environmental permits, especially on the backup generators, those can be challenging. And there are ways that you can manage that. I mean there are different quality of generator efficiencies and quality of emission controls. So certain generators are easier to get permitted, certain generators are more difficult. Really, it depends on how much you expect to use the generators and the associated emissions over the year. So the data center project can have the same backup generator, but based on what its expected uptime, could have 2 very different permits. So it's a bit of a complex and nuanced situation.

But we're always striving to find the ways to speed up and compress those time lines, especially if it's something like permitting. So we do evaluate all of the solutions out there with regards to BESS or different generator solutions to try and keep that process as quick and as efficient as possible.

Gregory Lewis: Okay. Great. And then I was hoping, Ben, you can talk a little bit about Sherbrooke. I guess just now that the power has, I guess, been across the site or however that's used by potential customers, [indiscernible] I guess, the 9,600 megawatts [indiscernible] plus 1 site. How does that -- what does that actually mean from a marketing perspective for Keel?

Benjamin Gagnon: Yes, that's a great question, Greg, and I'm happy to speak about the Sherbrooke project. So we've got a decent-sized portfolio in Quebec, and Quebec represents a market that is very captive. There's a lot of legislation in both Canada as a country and Quebec as a province that really strongly incentivizes data sovereignty at the national and at the provincial level. But unfortunately, it's just been very, very hard to secure new electrical capacity for data centers. What we have in the province of Quebec is we've got a huge energy portfolio, but specifically approved for Bitcoin mining.

And what the approval that we received on Sherbrooke was for consolidating 3 different Bitcoin mining power purchase agreements we have into a new single power purchase agreement, specifically for HPC and AI. And that one piece there, the change in the industrial use case is the big change here that enables us to actually move forward with developing an HPC and AI data center once we have the last sign off from the provincial minister.

And the reality is, is that because the legislation is there and because the demand is captive, we think that Canada and Quebec largely can charge a little bit of a premium on the exact same compute because they just are that much more captive and the capacity is just that much more scarce.

Operator: And the next question will come from Mike Grondahl with Northland Capital Markets.

Logan Hennen: This is Logan on for Mike. Ben, first, can you provide a formal update if Keel is still targeting 3 leases announced in 2026, given the extended time line now for Panther Creek? And maybe just an update on how demand has evolved over the last 90 days since that target was announced?

Benjamin Gagnon: Yes, happy to cover that, Logan. We're still in active due diligence and negotiations at all 3 of our sites. I think the commercial process is going incredibly well. At every 3 of our priority sites, we've got a lot of very interesting and sticky potential tenants who are working through the negotiation process. And I think at this time, while we're working through the negotiations, we're just going to continue to focus on working through those negotiations and the multiple parties as trying to give a clue or an indication as to where any particular negotiation for any particular site or tenant is at.

But we remain incredibly optimistic and confident based on the commercial process so far, based on the continued process that we have with permitting across all 3 sites as well as the other background works with the engineers, the supply chains, the fiber contracts, everything is continuing to move forward. And I think the closer you get to energization date, the more valuable your energy becomes by the day. And so it becomes an easier and easier commercial process when you're working through a 2027 delivery date as opposed to a '28 or 2029. And so that continues to keep us incredibly confident, optimistic and it also helps to keep our potential tenants very engaged.

Logan Hennen: Great. Yes, I appreciate the insight there. Then one more from us. Can you kind of formally update us on the Scrubgrass site, where that's at today, how that site is progressing and the demand you're seeing for that 2028 plus power?

Benjamin Gagnon: Yes, sure. Happy to give an update on Scrubgrass, although there isn't much of a substantive update to give. Scrubgrass is what we call a pipeline site. So this is a very exciting 1 gigawatt plus campus in Western Pennsylvania. But right now, the process for Scrubgrass is really in the energy application stage. So we have been working with the local utility there for a detailed load study for 750 megawatts. And we've also been working on the pipeline and engineering, as many investors know, for a pipeline to support 550 megawatts of on-site generation with CCGTs and an IPP who would come in and deploy the turbines, finance, operate and sell the power to the end customer.

At this stage, we are still working on securing the power. And until we have secured the power, and we have a firm final understanding of how big the site is going to be by what time, we're not doing the engineering work for building out the data centers or planning out the data centers. We have not submitted any permits or any proposals at this time.

We're really focused on securing the power and working through what we call a mass grading plan and kind of a site campus layout plan so that we can know, as soon as we get the power approved, where we're going to want to build buildings, how we want to build buildings, the size of the buildings, the number of the buildings, the cadence and that sort of thing. But at this stage, it's still really in the energy application phase, and we should be able to provide investors an update as early as December or January.

Operator: And our next question is going to come from Michael Donovan with Compass Point.

Michael Donovan: On Sharon, I was hoping we could discuss the cadence for RFS. Are you still expecting 30 megawatts for the first data center and then expanding it by the 80 megawatts?

Benjamin Gagnon: So we've been working on that, Mike, and we've been working on how do we compress our time lines as much as possible. And also how do we improve our power density. As of right now, we haven't updated it, but we are looking at ways that we can compress it into 1, 110-megawatt phase.

Michael Donovan: Okay. That's helpful. And then at Moses Lake, is an additional 10 megawatts at the site still an option?

Benjamin Gagnon: No, we've decided to give up that option, and we are just focusing on the 18 megawatts in Moses Lake at this time, and we have given up the option.

Operator: And our next question is going to come from Bill Papanastasiou with Chardan.

Bill Papanastasiou: Can we please double-click on the environmental permitting process. Are you seeing a higher bar being set given the recent political headwinds on building data center capacity? And more specifically, how would you assess the likelihood of environmental permitting approvals today relative to prior quarters?

Benjamin Gagnon: Thanks, Bill. Yes, happy to dig into that a little bit. I mean, really, when you look at our permits across both Sharon and Panther Creek, they're really kind of the same permits at both facilities. They're all environmental. It's largely associated with sewage, which is a pretty standard permit to apply for and get. It's not one that tends to be controversial as well as kind of the ground stuff. So things that deal with erosion, water, storm water is basically what the rest of the permits entail. So these are engineering focused. Like I said, they're not generally politically sensitive or subject to a whole lot of opinion. It's really just the engineering work.

And one of the things that we've mentioned on previous calls, I think people have asked about our relationship with the OTO, which is Pennsylvania's Fast Track Office, so Governor Shapiro has a fast track office for permitting. That's actually run out of the DEP because the DEP is well known for kind of taking the longest line item in the permitting process. And that's actually split up into 2 departments. There's a Eastern DEP and there's a Western DEP, and it's the Eastern DEP that tends to be the one that's a bit more overworked and it takes a bit longer to go through the permitting process than the Western one.

And so it's really just a matter of working through the backlogs. But this is a well-known -- these are pretty standard permits. This is a well-known process, and we remain incredibly confident, most confident we've ever been on completing our permits for both Panther Creek and Sharon today.

Bill Papanastasiou: Appreciate that. And apologies if this was mentioned, Ben, but the conditional approval at Panther Creek, what are the conditions attached to that?

Benjamin Gagnon: There's too many conditions to name, but to give you kind of like some examples, conditional approval will include things like you need to adjust your setbacks or maybe you need to adjust the height from 62 down to 60 or just something like that. They're pretty standard recommendations. It will be very specific. They will usually be very numeric, and it's make the following recommendations or implement the following systems or achieve the following conditions. They are not hard to comply with. And the real advantage of having that conditional approval, it's a very clear checklist of everything that you need to do so that, that conditional goes away and you are just fully permitted.

And so it's a very clear prescription or recipe or however you want to think about it for getting there. If you're -- if they don't want to get you approved then they wouldn't be providing such a clear road map for that success.

Bill Papanastasiou: Understood. And then there was a prior question on Quebec. Can you talk about that opportunity? How ripe is the sovereign AI market in the province? And how do you see Keel capitalizing on that?

Benjamin Gagnon: Yes. So we've spoken with a number of different industry experts, especially in the province. We think that rates generally in Canada are higher than they are in the United States, but it's hard to put a firm figure on that. But generally speaking, they are higher. And what we see is that there's some nice diversification benefits for us as a company. We have the U.S./Canada diversification element. So there is the element where, in Canada, you don't have to worry about regulatory changes with regards to tariffs and all of those other items, which might impact the cost of a data center.

So we think that delivery in Canada could potentially be cheaper than in the United States, and we think the market could potentially be worth more than it is in the United States. The challenge with Canada is the same challenge we've always had with Canada. It's just a very hard market to grow in organically. And so if you're looking to achieve a 1 gigawatt growth in Canada, that's probably a very, very, very high hanging fruit and much higher hanging fruit than trying to achieve 1 gigawatt at a campus like Scrubgrass in Pennsylvania.

But for the power that we have, we believe that working through to get that approval, working forward to make sure that we have the clear path, all the permits, all the support needed and secured for us to develop a data center, we believe we can generate some pretty attractive yields in Canada.

Operator: [Operator Instructions] And our next question will come from Stephen Glagola with KBW.

Stephen Glagola: Ben, how should investors think about the significance of the August 20 Department of Environmental Protection meeting for Panther Creek, and what are the key decisions or milestones that need to come out of that meeting?

Benjamin Gagnon: It's very routine meeting. I don't think you should be thinking about this as a special or a unique thing. It's just another routine meeting.

Stephen Glagola: Okay. All right. That's good to know. And I guess a higher-level question for you would be, when you're evaluating prospective tenants, to what extent does your view of the long-term model or landscape influence your willingness to partner with a particular AI lab?

Benjamin Gagnon: Well, that's a very interesting question, one that we actually think about a lot because the market is changing quite quickly. Even just last week actually, we were talking about the entire company, Keel has adopted Claude for our enterprise AI solution, but a year ago, none of us were using Anthropic, we were all using ChatGPT and now that's completely changed. So I think that model -- or I think the market is going to continue to change and adjust. This is a market where the incentive is very high. There's a lot of people who want to push for the top, and we do expect it's going to continue to change.

We think that Anthropic has found a nice niche in the enterprise market, which is the one that we've always been identifying as the one that's really going to be driving this industry forward as opposed to retail is going to be the enterprise consumers and maybe they develop a little bit of a moat here. But we're going to try and stay as agnostic as possible with regards to the models because as we've said before, a lease is really not a trophy for a press release, it's a 15-year commitment. And the gap between a good one and a bad one is measured in hundreds of millions of dollars.

We're not in this for the company who can only pay their rents for 1 year, right? We're in this to find the companies who are going to be able to give us long-term contracted, predictable revenue for 1 to 2 decades.

Operator: And our next question is going to come from Nick Chiles with B. Riley Securities.

Nick Giles: A lot of good questions asked already. So I just wanted to zoom out and ask, Ben, what do you really see as some of the biggest risks at this point? It seems like you made some progress on the supply chain front, but curious if there's any kind of further mitigation you can do there?

Benjamin Gagnon: Thanks, Nick. I think the biggest risk at this point is probably just broader macro. The reality is, is that there's still very little 2027 power that's available in the market. And we have a really strong position because we have a very reasonable amount of the 2027 leasable capacity remaining. So I think, broadly speaking, that keeps a lot of our -- well actually keeps all of our potential tenants incredibly engaged. It keeps them incredibly sticky. They're all looking to solve the exact same deployment problems. And so there's a real strong advantage there towards having that 2027 power that everyone is so focused on delivering.

I think the broader market is probably what we're watching the most, how our capital markets evolving and changing, how are the financing opportunities for the market changing, what's happening with interest rates and broader risk-on, risk-off sentiment, how is the market processing, the increasing amount of intercompany financing that we're seeing across the industry? I think those are the things that were really the bigger risk factors for the business. And fortunately, those are things that the entire industry kind of equally faces together. But given we have that 2027 power that's in high demand, we're incredibly highly confident with our portfolio and moving forward with the commercial processes for all of our sites.

Nick Giles: Great. And that's good to hear. And then just maybe on the CapEx side, I was curious if you kind of have any rough sense for where that could shake out? And if there's a kind of development cost or a certain yield to cost hurdle that you're looking to achieve on any signing?

Jonathan Mir: This is Jonathan. Thanks for the question and good to talk to you this morning. We continue to suggest that you use the rule of thumb industry averages that you might see in equity research for purposes of your own modeling in terms of construction costs and yield on costs and that should work well for you.

Nick Giles: Fair enough. I appreciate that, Jonathan. And then just one more, if I could. I think all the BTC sites have been decommissioned now. So should we really be zeroing out revenues for the balance of the year?

Jonathan Mir: So at the beginning of the year, we made clear from our liquidity forecast that we were assuming there would be no cash contribution from BTC embedded in any of our forecasting. We still do have rigs up in Canada as a practical matter. They might contribute 2 or 3 Bitcoin a day. But again, all of our discussion around liquidity and projected liquidity assumes that the Bitcoin business provides no cash over the course of the year.

Operator: And the next question is going to come from Martin Toner with ATB.

Martin Toner: Congrats on the progress. A question about timing. Now that Panther Creek, which is the flagship or crown jewel asset, not to put words in your mouth, is delayed relative -- likely relative to the others. How do you think that changes timing for deal announcements? I mean, is it possible a tenant wants all 3, and therefore, it will take a little bit longer to sign it? Or which one do you think will go first?

Benjamin Gagnon: Thanks, Martin. It's -- we have an internal betting pool in terms of which site is going to go first, but it's really, really hard to pinpoint exactly where that's going to land out. You don't really know what's going on in the background with each customer. And generally speaking, they're going to be as aggressive as their back pipeline of demand is there. So they're going to be quite aggressive depending on what's unique to them. With regards to a timing for Moses, Sharon, Panther Creek and whether or not 1 potential tenant could be interested in all 3, I mean, I can confirm that we have multiple tenants who want all 3 sites.

But that doesn't mean that's how we want to run the process or that's how we necessarily want to be building our portfolio. We'd rather be looking at trying to keep tenants focusing on individual sites, get them focused on one site that they can take down and then look at how they can build potentially a pipeline of growth with us beyond that first asset. So many of the things that -- many of the tenants that we've been speaking to recently are not just interested in an asset, they're interested in finding a development partner that they can continue to scale with predictably over time.

And so that's how a lot of these conversations are going is how do we get on with Moses Lake first, but then how do we also sign up for a second site or continued expansion in '28 and '29 with you. Same thing with Sharon and same thing with Panther Creek. Whether or not that impacts the timing for Panther Creek is not certain right now because the commercial process is still incredibly active and nobody seems to be batting an eye. As long as our RFS date remains 2027, I don't think there will be any impact here on our commercial process.

Martin Toner: That's very helpful. Has the RFS date within 2027 changed for any of the sites?

Benjamin Gagnon: For Panther Creek, we've always been end of year '27 and same year for Sharon, end of year '27. So I don't believe that we'd push back our Pennsylvania sites. I think Moses Lake has been delayed maybe a couple of months since our original guidance, but it's still going to be the first site that we expect to have online in next year.

Operator: And the next question is going to come from Brian Dobson with Clear Street.

Brian Dobson: I guess as you're looking at your portfolio, where would you like to add additional resources or expand in existing ones? I suppose, are there certain geographies that you're favoring more than others at this point?

Benjamin Gagnon: Thanks, Brian. That's a great question because we are looking at how we continue to grow our pipeline beyond '27 through '28, '29 and 2030. We do still have a global view, but we do have a strong, I think, preference for the East Coast, specifically the U.S. Northeast and the Midwest areas. We think those areas have tremendous energy resources and tremendous inference potential over the next couple of years and is going to be likely the areas where we see the greatest opportunities for HPC and AI infrastructure build-out. But it's early days. There can always be amazing opportunities outside of those areas, and we're certainly not going to be closed off to those amazing opportunities.

But I think, generally speaking, that's going to be where we focus.

Brian Dobson: Great. And then I guess in recent weeks, you've seen governors from New York and Texas, I guess, draw -- put an increased level of scrutiny on data centers. Do you think that this is something that we might start to see in other important energy regions? And ultimately, do you think it favors established players like yourselves?

Benjamin Gagnon: Thanks, Brian. The trend right now or the winds right now indicate we probably are going to see a few more headlines like this in the coming months. I think Pennsylvania represents a really, really unique centrist state in our view. This is a state that is very, very focused on energy and heavy industry. It's very, very blue in the major cities, and it's very red everywhere else. And so when you look at Governor Shapiro and kind of the politics of Pennsylvania, they do represent a very unique kind of centrist position for the United States these days.

It is one of the least polarizing states in my view in terms of the politics because they do know that they need to balance out the trades, the industry, the energy, all of those sort of employment opportunities, which is what drives Pennsylvania with the other concerns on the other side of the hall. So we think that this is a great place to be is in Pennsylvania.

We think that if states want to block themselves off from the best economic opportunity for development in decades and could be for the next couple of decades than we think that's pretty shortsighted because when you look at what one of these data center investments does for communities, for revenues, for employment opportunities, for tax budgets, for the schools and for the roads and what have you, these are transformative for the communities that we're investing in. And we think that they're very, very excited about the projects because of those investments because somebody is actually looking to do that. So we think it's pretty shortsighted, but we'll probably continue to see a few more.

And generally speaking, we think Pennsylvania is in a sweet spot.

Operator: I am showing no further questions at this time. I will now turn the call back over to Ben for closing remarks.

Benjamin Gagnon: Thank you all for joining us today, and thank you to the entire Keel team whose work this quarter speaks louder than anything I've said on the call. We'll see you all in November with more to show you. Thank you.