Keel Infrastructure Corp. (BITF) Earnings Call Transcript & Summary

Good day, and thank you for standing by. Welcome to the Bitfarms Fourth Quarter 2024 Earnings Conference Call. [Operator Instructions] Please be advised that today's conference is being recorded. [Operator Instructions] I would now like to hand the conference over to your speaker today, Tracy Krumme, SVP of Investor Relations and Corporate Communications.

Thank you. Good morning, everyone, and welcome to Bitfarms Fourth Quarter 2024 Conference Call. With me on the call today is Ben Gagnon, Chief Executive Officer and Director; and Jeff Lucas, Chief Financial Officer. Before we begin, please note that this call is being webcast with an accompanying slide presentation. Today's press release and our presentation can be accessed at our website, bitfarms.com, under the Investors section. Turning to Slide 2. I'd like to remind everyone that certain forward-looking statements will be made during the 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, and I invite you to consult Bitfarms' MD&A for a complete list. Please note that references will be made to certain measures not recognized under IFRS 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 MD&A for definitions of the aforementioned non-IFRS measures and their reconciliations to IFRS 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, Chief Executive Officer and Director. Ben, please go ahead.

Thanks, Tracy, and thank you, everyone, for joining us today. On today's call, I will be discussing our 2024 and year-to-date 2025 accomplishments and providing an overview of what the new Bitfarms is focused on in 2025 and beyond. Turning to Slide 4. Bitfarms is a completely different company than we were at the beginning of 2024. Across nearly every dimension, we have rapidly transformed ourselves from the international Bitcoin miner to a North American energy and compute company. The improvements speak for themselves. In our Bitcoin compute portfolio since January 1, 2024, we've nearly tripled our hashrate to 18.6 exahash under management and improved our efficiency by 45%, reaching our Q2 2025 efficiency guidance of 19 watts per terahash, 3 months ahead of schedule. Both of these improvements were driven by our fleet upgrade, which reduced the average age of our active miner fleet from approximately 3 years old to approximately 1 years old today. In our energy portfolio, over the same period, we've increased our energy capacity by over 90% to 461 energized megawatts today. We've built an energy pipeline of over 1.4 gigawatts with nearly 80% based in the U.S. And we reduced our expected average price power to $0.043 per kilowatt hour and achieved far greater control over our energy costs through our PJM portfolio. With higher efficiencies in our compute portfolio and lower energy costs in our energy portfolio, we have driven nearly a 50% reduction in our hash cost to roughly $22 per petahash. This largely derisks our portfolio and provides a foundation of higher levels of profitability and cash flow through 2026 and beyond, while we focus the company and our capital on new growth opportunities in the U.S. and with HPC. Importantly, Bitfarms is now a U.S.-focused company with the vast majority of our pipeline in the U.S. for the first time in our 8-year history. The U.S. now represents 33% of the energized megawatts in our portfolio, and this is expected to increase to nearly 80% over the coming years as we execute on our growth pipeline, almost all of which is in the U.S. Our momentum has only accelerated during the first quarter. We've closed our transformative acquisition of Stronghold Digital Mining, the largest M&A deal between 2 public miners in our industry. We closed the strategic sale of our 200-megawatt Yguazu data center, our largest site ever constructed. We advanced our HPC and AI strategy with the engagement of 2 new advisers. We hired 2 new critical team members, an SVP of Infrastructure and an SVP of HPC and AI, and we significantly improved our hashrate, reaching 18.6 exahash under management, which if ran at 100% 24/7 would be capable of producing over 10 Bitcoins a day with current mining economics and which we expect to be competitive in generating free cash flow through 2026 and beyond. We are a very different and a much stronger company today than when we started or even ended last year. And I would now like to spend some time discussing how our 2 most recent transactions were such big drivers in our transformation and why they have set us up so strongly for an exciting future in 2025 and beyond. Moving to Slide 5. Just last week, we announced both the close of our acquisition of Stronghold Digital Mining and the close of the strategic sale of our Yguazu site in Paraguay. While these are 2 separate transactions, they both fall under the same portfolio management strategy, and so it is helpful to view their impact together. Completing these 2 transactions have completely transformed our portfolio, and we are now bigger, better, and stronger as a result. First, our energy portfolio is both larger today and able to grow bigger and faster into the future. Inclusive of the sale of our largest ever build, Yguazu, we still increased our energized megawatts and secured a growth pipeline of energy assets that could scale our Pennsylvania infrastructure portfolio to over a gigawatt in the coming years. Second, we are driving stronger operating economics across our portfolio. By switching to more cost-effective American energy and by managing our energy costs through the use of PJM's robust power trading, curtailment, and demand response programs, we are able to reduce our #1 operating expense, energy. Controlling our energy cost is critical to controlling our future, and we are now able to do so with much greater scale, precision, and flexibility with our growing PJM portfolio. Lastly, both of these accretive transactions have better capitalized us for 2025 and beyond for the following reasons. First, the sale of Yguazu reduced our planned CapEx requirements for 2025 to less than $100 million, 20% lower than we laid out at the start of the year. And frees up the invested capital and profits from the sale of our Iguazu site for reinvestment in U.S. infrastructure. Second, we've avoided the additional significant expenses that would have been required to both complete construction, energize, and fill the 200-megawatt Iguazu site with new miners. We estimate this to be a savings of $325 million. And third, with Stronghold, we acquired 2 fully operating sites, both of which have been upgraded over the last 4 months with our hosting agreements. This transaction was primarily paid for in equity, and all Stronghold loans have been paid in full, saving significant interest expense and preserving our financial flexibility to work with strategic financing partners for potential HPC and AI projects. To summarize, the strategic rationale for both buying Stronghold and simultaneously selling Yguazu are 2 sides of the same coin. Through these transactions, we have rebalanced our portfolio to the U.S., where we expect to achieve greater yields per megawatt. We have reduced our average cost of power across our portfolio. We have minimized our CapEx requirements while improving both liquidity and direct operating margins to help finance new infrastructure investments. And lastly, we have secured what we believe to be highly desirable sites for HPC that enable us to diversify beyond Bitcoin mining into energy generation, trading, and HPC and AI. This level of portfolio management is unheard of in our industry. Closing a large public M&A deal simultaneously with a large strategic sale was a heavy lift, and I'm incredibly proud of the Bitfarms' team and what we have accomplished here in this quarter. We have transformed this company in short order, and it is truly a testament to how much stronger the new Bitfarms is today and what we can accomplish in the future. Turning to Slide 6. As we look towards our future, it now looks very different than it did a year ago today. With our vertical integration into power generation and having secured the largest growth pipeline in the company's history, we now need to think about our megawatts a little differently than we have in the past. I believe the simplest way to think about it is in 3 buckets. The first bucket consists of 461 energized megawatts and tracks megawatts currently being monetized. Today, this primarily means via Bitcoin mining. But as we diversify beyond Bitcoin mining, it will eventually include other use cases like HPC and AI. For investors who are interested in Bitfarms' current operating footprint and current cash flow out of operations, this would be the most important metric. The second bucket consists of over 131 megawatts under development and measures megawatts that are secured for near to medium-term growth. These are megawatts that are approved for development and don't require any further regulatory approvals or permits to begin development or energize. For investors who are interested in near-term growth potential, this metric best captures it. The third bucket consists of over 800 megawatts in the megawatt pipeline and measures the megawatts that are in process, but still require more regulatory work, approvals, certainty or clarity before they can be developed over a multiyear time frame. For long-term investors who are interested in our growth potential through 2028, this metric best represents the company's long-term potential, direction, and strategy. Turning to Slide 7. This slide shows the path to over 1.4 gigawatts of power capacity and how our growth efforts transform our portfolio into one comprised of nearly 80% U.S. megawatts and 90% North American megawatts. As you will see, we have 461 megawatts energized today, growing to 500 megawatts by year-end, the results of planned megawatts under development at both Sharon and Panther Creek. Likewise, in 2026, we anticipate growth of 92 megawatts for a total of 592 megawatts coming from Sharon, Panther Creek, and Washington. Finally, growth from year-end 2026 through 2028 is anticipated to come from megawatts currently under study and application at Panther Creek and Scrubgrass for a total of 1.4 gigawatts. As we look to this future and execute on our growth pipeline, we will continue to improve our portfolio composition by prioritizing the U.S., Canada, and those energy assets that have opportunities beyond Bitcoin. We expect this will drive both lower overall energy costs and better yields per megawatt, creating lasting value for shareholders. Turning to Slide 8. The largest and most exciting part of our megawatt pipeline is the gigawatt of potential in the U.S. Our new focus on the U.S. and our American energy assets are attracting a lot of interest from investors. Through these assets, we are opening up exciting new opportunities for Bitfarms in areas beyond Bitcoin mining, including energy generation, trading, and HPC and AI that weren't possible in Latin America and will define this company's future for many years to come. This gigawatt of potential is easiest to understand when separated into the tranches discussed on Slide 6 and separating out the generation as we have done here on this slide. At both Panther Creek and Scrubgrass, we have 2 operating power plants that are capable of producing power and selling power into the grid. They are also able to draw down and consume from the grid a similar quantum of power that the power plants can produce. As of today, it's not possible to monetize 100% of the power generated from the power plants and the power from the grid simultaneously. To do so, we will need to make further investments into the data center infrastructure and get greater clarity from FERC. We have strong confidence that this regulatory clarity is coming and will support our site plans. And while this regulatory approval process is ongoing, we can choose to provide 142 megawatts of power for the existing data centers from either the power plants or the grid at our discretion. Currently, our data centers are being supplied by the power plant as the lower cost option. Beyond this, we also have 6 different power applications across the 2 sites that are well underway and at various stages in the approval process, but cumulatively have the potential to scale both of these sites between 400 and 500 megawatts by 2028. Lastly, with 120 megawatts of potential, Sharon is the smallest Pennsylvania site, but still bigger than any of our sites outside of Pennsylvania. Sharon has 12 megawatts operating today, increasing to 30 megawatts in the second quarter. Sharon also has 80 megawatts actively under development for 2026 and 10 megawatts in the pipeline that is subject to additional reviews, studies, and approvals for a total of 120 megawatts. The 3 sites add up to a cumulative 1 gigawatt of potential. Turning to Slide 9. I'd like to take a minute now to touch on our HPC and AI strategy in more detail and discuss a few key strategic moves we've made in the last few weeks that have positioned us very well to capitalize on this massive growth opportunity. Just last week, 7 members from our team participated in the annual NVIDIA GTC Conference in San Jose. Our team was booked back-to-back with potential partners in the space from all angles, including financing, development, infrastructure, and end-use customers. Most of the interest was focused on 3 things: Power, land, and fiber, and that's what we have in spades at all 3 Pennsylvania sites. Having met one-on-one with over a dozen respected counterparties in the space, the feedback was all incredibly positive and exciting. Each site has power availability in the short-term, long-term growth potential, and multiple sources of power, including on-site generation at Scrubgrass and Panther Creek. Having multiple sources of energy improves reliability and reduces anticipated CapEx and OpEx for HPC and makes these sites very appealing for potential HPC customers, especially hyperscalers. This message was reinforced in Jensen Huang's address at GTC as power opportunities are in short supply and are the primary bottleneck on growth. At Bitfarms, we've known this for a long period of time, and it's one of the many reasons why our team was so bullish on the Pennsylvania sites. On the back of our pivot into HPC, we've also made 2 critical hires, James Bond, our new SVP of HPC; and Craig Hibbard, our new SVP of Infrastructure. James is a subject matter expert in HPC and AI with over 20 years of experience and joins us from Hewlett-Packard, where he spent 15 years leading the North American HPC and AI infrastructure platforms category, which grew to $2 billion in 2024, representing 160% year-over-year growth. At HPE North America, James was responsible for all HPC and AI go-to-market activities and his experience implementing HPC solutions at scale make him the ideal candidate to lead this new chapter at Bitfarms. Craig, our new SVP of Infrastructure, has extensive experience leading large-scale digital infrastructure projects and recently spearheaded the rapid design and construction of over 200 megawatts of digital infrastructure for a U.S. firm specializing in digital assets in HPC and AI. Based in Pennsylvania, Craig will play a critical role in leading the development of our rapidly expanding PJM portfolio. James and Craig will be working intimately with our strategic partners, ASG and Worldwide Technology to complete this comprehensive feasibility analysis on all of our North American sites and advise on our global HPC and AI strategy. In parallel, ASG and WWT will help build accelerated sales and development strategies and market our sites on behalf of Bitfarms to potential customers. Both of these initiatives are well underway, and ASG has already delivered initial high-level feasibility reports for our 3 PJM sites. All 3 reports confirm that the Pennsylvania sites are both well suited for HPC and AI conversion and high potential. All 3 are strategically located near other data center campuses and peering hubs, and they have the necessary power, land, and fiber infrastructure to support HPC. The cooler climate in Pennsylvania should also be expected to improve power efficiency and cost, which for HPC makes Pennsylvania megawatts more valuable per unit than the same megawatts in Texas. These 3 maps show the fiber connectivity at our 3 Pennsylvania sites: Panther Creek, Scrubgrass and Sharon. Each site has robust fiber infrastructure within 10 miles of each facility. Fiber infrastructure is one of, if not the first question asked when meeting with hyperscalers and other potential partners. Scrubgrass has 7 metro networks, Panther Creek has 8, and Sharon has 9 in close proximity. It's also important to note the proximity of these PJM facilities to metro areas and long-haul infrastructure like peering and interconnection hubs. Scrubgrass and Sharon are both approximately 1 hour from Pittsburgh and Panther Creek is just 2 hours away from both New York City and Philadelphia. We expect to receive full detailed feasibility studies from our partners in Q2 for our remaining North American sites, at which point, we will provide an update regarding our plans with HPC and AI. With these recent actions, we now have the properties, the internal team and the strategic engineering and marketing advisers in place, taking a holistic approach to advancing our HPC and AI business. Turning to Slide 10. Shifting gears a bit, I'll now discuss a new strategy we recently initiated, Bitcoin One. This is a continuation with a few strategic changes of our Synthetic HODL program, which was launched as a pilot program in Q4 of 2023. As a reminder, that strategy involves selling Bitcoin to fund OpEx and CapEx while preserving upside to Bitcoin through long-dated Bitcoin call options for between 10% and 20% of each Bitcoin sold. This program enabled us to utilize excess Bitcoin generated each month to fund our growth at a far lower cost of capital than external or dilutive funding sources while maintaining upside to Bitcoin. This program was a key source of low-cost capital, enabling superior returns and outperformance to Bitcoin. Since the program's inception in Q4 of 2023 through December 31, 2024, it outperformed Bitcoin price, and we achieved 135% return in U.S. dollar terms, generating a trading profit of approximately $18 million. Building off this proven success, we developed a more robust program, which we initiated in Q1, Bitcoin One. This is a quantitative investment multi-strategy program that employs a disciplined approach to Bitcoin accumulation through diversification, strategic leverage, and market timing. Bitcoin One focuses on active Bitcoin treasury management through discretionary and rules-based trading algorithms and an active managed volatility targeting program that trades crypto volatility as an asset class and harvests the risk premium that arises from that volatility. This program is positioned for the ongoing maturation of the digital asset ecosystem, providing a scalable and risk-managed solution for Bitfarms to monetize Bitcoin's volatility. We believe that the combination of Bitcoin One and the predictable cash flow of Bitcoin from our mining operations gives us a unique advantage in both markets that enable us to drive better probabilistic returns and accumulate more Bitcoin faster and cheaper. Turning to Slide 11. In summary, we've made significant progress year-to-date and expect this momentum to continue throughout the rest of 2025. All of the strategic moves we've made in the last few months have been carefully considered and made with 2 main goals in mind: To grow our operations in the U.S. and to utilize our megawatts in the most value-add and cost-effective way possible, whether that be through Bitcoin mining or HPC and AI. Our key strength is acquiring and managing valuable property assets and maximizing the value of those assets through our operational excellence. I'd now like to highlight 4 main takeaways before I pass the call over to Jeff. First, we are now in an extremely strong position. We've completely transformed our growth pipeline over the past 15 months from Latin American energy to a cheaper, bigger, and better U.S.-based pipeline. Second, our Bitcoin business is strong, and we remain bullish on mining economics for the mining fleet that we have. With our lower hash cost derived from our competitive energy prices and greatly improved efficiency from the fleet upgrade program, we anticipate that our Bitcoin compute portfolio largely derisk the business through 2026 and should provide a strong cash flow and low CapEx foundation for the company as we look to diversify beyond Bitcoin mining. Importantly, with this new competitive fleet that we now operate, we have no need nor plans for a large miner purchase in 2025 or 2026 and will instead focus our efforts and capital on developing U.S. energy and HPC infrastructure. Third, we are taking a long-term view on how to best manage our energy assets. While we remain bullish on Bitcoin mining and are confident in our core business, especially in the near-term, in the long-term, we believe HPC and AI presents a meaningful opportunity to create lasting shareholder value, which is why we have taken tangible steps to advance our HPC business. We have now secured the properties, hired the best internal team, and brought in the strongest external partners. Lastly, I'd like to emphasize the importance of the significant changes we've made to our team over the past few months. With the recent hiring of our SVP of HPC and AI and our SVP of Infrastructure, we have now filled all key management positions in the company. We've also completely revamped our operational structure. And as evidenced by the accomplishments in Q1, we are now beginning to reap the benefits of the strong team we worked incredibly hard to build. I'd like to thank all of our team members for their commitment to the Bitfarms' mission, and I've never been more confident in our team and our ability to execute. Turning to Slide 12. And with that, I will turn the call over to Jeff for the financial update.

Thank you, Ben, and thanks, everyone, for joining us today. Before I review our fourth quarter results, I'd like to underscore just how transformational the initiatives Ben reviewed will be for our company. First, we are better capitalized for 2025 and beyond. The Stronghold transaction was paid primarily with equity and its outstanding debt of approximately $45 million was paid in full. The sale of Yguazu reduced our planned CapEx requirements for 2025 to less than $100 million, 20% lower than originally planned, freeing up invested capital from the sale for reinvestment in U.S. infrastructure. Second, our Bitcoin mining business remains strong. With the lower hash cost of our new fleet, our Bitcoin mining portfolio is largely derisked and positions us to build from any potential BTC upside. Importantly, we have no plans for large miner purchases in 2025 or 2026 and will instead focus our capital on developing U.S. energy and HPC infrastructure. Third, our highly valued North American assets and the cash flow from our mining operations will enable us to utilize debt financing to fund our energy and HPC infrastructure development at a lower cost of capital and with significantly less dilution than equity funding, creating long-term shareholder value. Please now turn to Slide 13. Let's review our fourth quarter financial results. In the fourth quarter, we earned 654 Bitcoin for total revenue of $56 million, up 21% compared to the third quarter. Revenue from our mining activities was $55 million with the balance of $1.6 million earned at Volta, our electrical services subsidiary. Our direct mining profit was $26 million, representing a direct mining margin of 47% and an average of $39,000 per Bitcoin mined. Cash, general, and administrative expense, or G&A, was $14 million and included unusual and nonrecurring professional service fees of about $1.3 million in the quarter, primarily for expenses incurred in connection with the Stronghold acquisition. Operating loss was $16 million in the quarter. And for the quarter, the company reported net financial income of $22 million, inclusive of an $18 million gain on derivative assets and liabilities, largely driven by our Synthetic HODL program and a $6 million gain on the revaluation of warrant liability for warrants issued in earlier financings. As a result, net income for the fourth quarter was $15 million or $0.03 per share compared to a net loss of $37 million or $0.08 per share in the third quarter. Turning now to Slide 14. For the fourth quarter, our adjusted EBITDA was $14 million or 25% of revenue compared to $6 million or 12% of revenue in the third quarter. As we've noted in previous quarterly earnings calls, our adjusted EBITDA is very straightforward, being purely a measure of the cash profitability of our mining operations and the profit contribution of our Volta subsidiary. We do not reflect mark-to-market adjustments of our Bitcoin holdings or any other balance sheet valuation adjustments in our adjusted EBITDA. Now let's turn our attention to fourth quarter operating performance and per Bitcoin metrics. Gross mining profit in the quarter was $26 million or 47% of mining revenue. Our direct mining cost per Bitcoin in the fourth quarter was $40,800 with all-in cash cost to mine a Bitcoin at $60,800 compared to revenue per Bitcoin earned at $83,400, resulting in cash profit per Bitcoin of just over $22,600. Turning now to Slide 15. I'll review our liquidity and anticipated capital needs. As of March 26, we had total liquidity of approximately $135 million, comprised of cash and unencumbered Bitcoin. In addition, we have $31 million in payments from HIVE due over the next 6 months from the Iguazu sale, and our mining business remains strong, generating $8 million to $12 million per month in cash from our mining activities, providing more than sufficient liquidity to fund our identified CapEx for 2025. I wish to emphasize a few important points about our financial requirements going forward. First, 2024 represented a one-off fleet-wide upgrade and investments into new miners. Second, the strategic sale of our 200-megawatt Yguazu facility reduced our 2025 CapEx requirements significantly by over $325 million, including $85 million on the build-out of the facility and an additional $240 million or so for procuring and installing miners. Third, with the Yguazu sale done and our fleet upgrade largely complete, the company has no identified plans for large miner purchases in 2025 or 2026. So for the balance of 2025, identified capital expenditures, including infrastructure development and the remaining deployment of our miner upgrades are projected to be under $100 million with the lion share of that capital being invested in infrastructure to increase our megawatts under management from 460 megawatts today to 590 megawatts in 2026. Lastly, with virtually no debt on the balance sheet and strong cash flows from our mining business, we believe we have a lot of options to finance potential HPC and AI projects and are engaged in active discussions for development of our North American sites. Turning now to Slide 16. Before we proceed to Q&A, I'd like to echo Ben's enthusiasm about the strong position Bitfarms is in with some market commentary on Bitcoin mining economics. While we recognize that Bitcoin price pulled back during the first quarter, we believe the current mining sector represents a potential opportunity for investors of most publicly traded miners, especially Bitfarms, and here's why. This hash price sensitivity table shows our best current estimate of mining economics for 2025 and 2026. On the Y-axis is Bitcoin price. On the X-axis is current network difficulty and our forecast of network hash rate at the end of every quarter through 2026. With these 2 variables, we can create a hash price sensitivity table. To make the chart more impactful, we overlay an implied direct mining margin relative to a hash cost of $22 per petahash and our best forecast on expected Bitcoin price ranges per quarter through 2026 as represented by the highlighted sections outlined in black. This table makes it clear that when combined with competitively priced electricity like Bitfarms has, new gen miners are largely insulated from Bitcoin price pullbacks and are expected to generate free cash flow in even challenging macro conditions during 2025 and 2026. Assuming Bitcoin at $80,000 today with 20 watts per terahash efficiency and $43 per megawatt hour electricity costs, a miner would achieve a direct margin of roughly 50% with substantial potential upside to rising Bitcoin prices. If Bitcoin is able to stay over $100,000 or $125,000 for most of 2025, direct mining margins should range between 50% and 75% for the year. This is an industry that could perform quite well in 2025. With that, I'll turn the call over to Ben before we commence Q&A.

Thanks, Jeff. Just like to reiterate for everyone that our original investment thesis on Bitcoin remains unchanged. We continue to believe that 2025 will be bullish for Bitcoin mining economics, and it has high potential to create a lot of shareholder value in 2025 from our existing fleet and footprint. We expect a probable correction in mining economics in 2026, driven by continued network hash rate growth and a correction in Bitcoin price from a new all-time high, which we expect to set later in 2025. This dovetails well with a more realistic expectation for bringing in HPC revenues in 2026. Even with a 2026 correction under these conditions, it would likely bring us back to similar levels of hash price economics and direct margins that we see today, which are healthy. This is the base case, and there is a lot of upside potential and exposure to rising Bitcoin prices, especially in 2025. Thank you. And with that, I will pass the call back to the moderator for Q&A.

[Operator Instructions] Our first question comes from Brett Knoblauch with Cantor Fitzgerald.

Congrats on the 2 new hires. It's nice to see. As maybe you look at the Stronghold sites, clearly, there's a plan for you guys to potentially use those for more AI/HPC capacity rather than Bitcoin mining. I guess can you just give us an update on timing there from a regulatory stance of what you guys need, when you expect to get those approvals? And to what extent the site as is today is of interest to potential partners without those approvals? Or do we need those approvals first?

Yes. So there's a couple of different buckets of energy there at the Stronghold site, as I spoke to on one of the earlier slides. For the existing generating capacity and the existing grid import capacity, there's no further regulatory approval to operate either of those. Right now, we can either provide power directly from the generating assets to the data center or we can draw down a similar quantum of power from the grid. Where regulatory approvals come into play is when we're trying to look at a configuration for all of the megawatts on those sites. So that's the additional megawatts that are under study and application, which can be as fast as maybe 12 months and as long as 36 months across the 6 different applications in the 2 different sites. But I think what's really interesting for the HPC [Technical Difficulty] is that we have immediate power capacity now in 2025 that [Technical Difficulty] for HPC, it would be available in the 2026 time frame, and this has a multiyear growth potential. And when a lot of these hyperscaler or HPC customers are looking to invest 100 megawatts [Technical Difficulty] thinking about is I'm going to make investment into site in a multiyear time frame. I want to know whether I'm going to be able to [Technical Difficulty].

And then, Jeff, maybe a question for you. Just on the CapEx side. I know you guys said you're expecting less than $100 million in CapEx for this year. Could you maybe just bucket that? I guess, how much of that is going towards Bitcoin mining versus maybe how much of that is going towards infrastructure that could potentially be leveraged or used for AI/HPC?

Sure. Very glad to do that here. So as I mentioned, we have less than $100 million of capital expenditures, more like around $95 million for the year. Really, Brett, only a small portion of that is going to be going towards miners, and that's really more logistics costs associated with implementing and completing out the upgrade program that we began actually in 2024. So around $7 million of the $95 million is going towards that. The balance of the money really is going towards building out the infrastructure, predominantly regarding Stronghold and the Sharon facilities here with a smaller portion actually happening in the Baie-Comeau in the Quebec area and some final work that's being done in Argentina as well.

Our next question comes from Mike Colonnese with H.C. Wainwright & Company.

Congrats on all the progress you guys have made over the recent months here. First one on HPC/AI side, and Ben, you alluded to time line on the approval and available energy. But it sounds like the early findings from the HPC/AI feasibility studies are going really well. And I was just curious, based on conversations you're having with prospective customers, be it hyperscalers or other AI parties, what are their due diligence timelines like? And how would a customer most likely utilize your energy assets at those facilities? Would it be more of a shell type setup, fully constructed data center? Just trying to get a better sense of the time line to monetization if a deal were to be secured over the near-term.

Yes, it's a great question. When you're looking at potential deals with hyperscalers, these are not Bitcoin miners. They're not acting with the same sort of kind of aggressive growth stance where you could get a deal done in a couple of days. These companies really spend a lot of time on their due diligence. And I think an aggressive time line would be 6 to 9 months. I think a more realistic time line is 9 to 12. And it could go beyond that as well. The conversations are such though with a hyperscaler that you are going to know fairly early on if the site is not of interest to them. They're not going to spend 6 months to do a quick evaluation. They're going to do a quick evaluation and see whether or not the site is worth investing time into. But to complete the full diligence, it's probably a 9 to 12 month estimate. Now when you're looking at how do you build out the time line and what does an actual structure look like. The opportunities here is that especially on the powered land and the powered shell components, those are things that you can start in advance of a deal taking place. And it's something that's going to be helpful for advancing the deal while the deal is going through the diligence process and also accelerate the timelines and the expected valuation of it because the time line to energization here is really what's driving value in HPC deals this year. The potential CapEx and potential returns that you get across these scenarios, you can think of it like a spectrum, and they're not really mutually exclusive. When you start on the lowest CapEx side, which is just powered land, CapEx requirements are pretty modest, and you can get a fairly decent return from a development perspective [Technical Difficulty] or the end customer lined up to start working on a powered shell. And then beyond that, you have a full build-to-suit. So really, you have a spectrum of options. From our perspective, we'd start on powered land and powered land construction could start in advance of the deal being signed. And it would be helpful for advancing a deal with a potential hyperscaler client.

And on the Bitcoin mining side, I know you guys laid out the 21 exahash target last year. Just curious as to when you'd expect to achieve that. I know you mentioned 18.6 currently operating. And then is there a potential to -- or is there a market environment in which you'd potentially expand beyond that number? I know you guys are really focused on developing the power assets and infrastructure. Just curious if, let's say, Bitcoin and hash prices were really to soar here if you'd have the willingness to go ahead and expand beyond the 21.

Yes, it's a good question. And we have been making a lot of progress here on our hash rate, especially this quarter. We're at 18.6 as of yesterday, and that number is continuing to track upwards. We're expecting to see continued progress here on our operational hash rate over the second quarter. The big gap here between where we are today and 21 exahash is largely due with the minor repairs, the RMAs and the upgrades that have been going forward. We recently signed a smaller upgrade deal. So we're upgrading about 4,000 of the miners, and that will be done in the second quarter. And it's really going to be a function of when the RMAs are completed and when these upgrades are done for us to be able to push that number as close to the 21 as possible. When it comes to future miner purchases, as I said on the call, we don't have any plans right now for future miner purchases. We're really happy and confident with the fleet that we have now. We think it positions us really well for 2025 and 2026. We do have opportunities here for potentially increasing our exahash under management if other companies or customers would like to pay for the miners through hosting, and we can achieve, I think, better returns on investment with potential hosting customers in Bitcoin mining. But our focus here is really on directing capital to source longer -- those more sticky long-term assets like the energy infrastructure and the HPC infrastructure. And we just think that's a better allocation of capital for 2025 and 2026.

Our next question comes from Michael Grondahl with Northland.

Ben, you said something in your prepared remarks, and I just want to make sure I heard it right. I think related to the Pennsylvania sites, you said 142 megawatts were currently available and then you needed that clarity from FERC over the next 12 to 36 months on the remaining. Is that correct, the 142 are available today across those 3 sites?

So let me break that down a little bit more specifically. So we have 142 megawatts of active data center capacity across Panther Creek and Scrubgrass today. And that 142 megawatts can be provided or can be powered up by either of the power plants or the grid. And that's done at our discretion, whichever is the more cost-effective option is the one that we can use, and we can switch back and forth as needed or as necessary over time. The expansion beyond 142 is really going to be dependent on 2 things. One, investing in more data center infrastructure so that we can consume or power more than 142 megawatts of compute at those 2 sites. There is more power available at both of those 2 sites to build out more data center compute in 2025 and 2026, and it's part of the plans there for us to reach 500 megawatts by the end of this year is to develop additional power and capacity at the Panther Creek site specifically. The -- where the regulatory approvals come into place is with regards to a configuration where we use the grid connection to provide a primary power source for the data centers. And then use the power plants to provide a redundant source of power to the data centers. And this is something that's going through the regulatory approval process with a lot of large-scale data centers. Amazon and Microsoft, I think, are the big ones who are driving this effort forward with FERC. And that's kind of what is going on there with the regulatory approval processes for running both of those simultaneously. And then with regards to the 6 power applications, that's less of a regulatory approval process and more of just a standard applications process. Power applications and these large interconnection agreements, they take time to work through the system and there's a bureaucracy in place there. That just takes time. So the fastest that we could expect, I think, any of these applications to be approved and move from the megawatt pipeline to the megawatts under development would be about 12 months. And on the long end for the full scale of those 6 applications, it's about 36 months.

And then you also mentioned a Q2 feasibility study. I think that related to other sites, not Panther Creek, Scrubgrass, and Sharon. But could you just clarify that too?

Yes. So we've engaged our 2 strategic advisers to help us run feasibility assessments across all of our North American sites. So they first prioritized the Pennsylvania sites, Panther Creek, Scrubgrass, and Sharon, but they're also doing Washington and our Quebec portfolio. So in the second quarter, we expect to receive all of these reports for all of the sites from both of the partners. And that's really going to be a crucial step and milestone here for us advancing our HPC and AI strategy. Because that's when we should have the full suite of information in hand for us to be moving forward more aggressively.

Our next question comes from Brian Kinstlinger with Alliance Global Partners.

Thanks for all the details on new capital allocation strategy. With the key hires in place, maybe I missed it, can you update us on the HPC pilots you talked about in November and maybe the timelines you expect for those projects?

So yes, it's a good question. We haven't made any announcement on the HPC pilot. And the reason for that is because we had one potential customer who is very interested in a small 1 to 2 megawatt pilot project, particularly at our Washington locations. And that pilot project was largely going to be driven by that customer demand. As we move forward with those conversations, their scope and their interest went from more of a pilot project to a full site potential. So increasing the capacity from about 1 to 2 megawatts to maybe 8 to 15 megawatts of potential capacity. But it's taking -- it's going to take more time in order to build out whether or not that's a viable alternative for those sites and whether or not they've got the financing on their end in order to move forward. So as of right now, we don't have any time line or any plan for the pilot and when that could be developed or energized. And we're going to be focusing our efforts and our capital on the powered land and the -- probably the powered shell potential here in the Pennsylvania sites.

Our next question comes from Nick Giles with B. Riley.

Ben, you noted the longer due diligence process for hyperscalers. I was wondering if you just expand a little more on customer preference. Should we think about hyperscalers at the very top? And would that require more of the build-to-suit model? Or -- yes, I'll start there.

Yes. It's going to be specific to the hyperscaler customer. Certain customers would or certain hyperscalers are much more interested in build-to-suit. Certain hyperscalers are much more interested in empowered land and doing everything all themselves. It really depends on the hyperscaler and their strategy. But I think the hyperscaler is the preferred way to go. It is a bit of a trade-off with a non-hyperscaler client, you could probably achieve higher margins, but you have a less strong credit with a counterparty. You're going to have higher cost of capital associated with financing a project with a non-hyperscaler tenant. And so when you look at the trade-off between margins and cost of capital on a project like that, I think the all-in blended figure favors the hyperscalers over another counterparty, but it's really going to be site and deal and counterparty specific. And so it's hard to say anything definitively one way or the other. I would say, in general, the creditworthiness of a hyperscaler customer and the lower cost of capital associated with financing a hyperscaler customer and the infrastructure for them is likely going to provide a better return and value for shareholders.

And you noted no large CapEx plans for Bitcoin mining in 2025, but you guys have made some pretty timely minor purchases in the past. So could we see you be opportunistic if machines were to trade at a discount? Or should we think about something like that is off the table at this point?

Yes. That's a really good question. The way that we approach miner purchases is, I think, pretty different than the way a lot of the industry does. We've never been focused on growth for growth's sake. We've always been a company that focuses on return on invested capital and trying to get a better deal for shareholders through not buying the most expensive miners or the most efficient miners and really focusing on those miners where we can get a little bit of an edge and a better return. With miner prices today, we still have not seen miner prices really adjust really at all over the last maybe 15, 18 months, and they stayed flat as Bitcoin has gone both up and down and mining economics has kind of fluctuated here. If economics change, it's certainly something that we would -- we could look at, but we don't expect that to happen. And we still think that over 2025 and 2026, our fleet is really competitively positioned from an efficiency and a cost perspective. We expect that the Bitcoin price and the way that the network hash rate is going to grow, those sites are going to be generating nice margins and free cash flow for quite some time. So there's not going to be any financial pressure for us to upgrade the fleet. We're going to just focus on optimizing the fleet running at the best that we can and achieving those diversification beyond Bitcoin objectives that we've outlined.

Our next question comes from Bill Papanastasiou with KBW.

Congrats on all the progress recently. For my first question, perhaps you can speak to the terms that you'd expect to secure from AI/HPC hosting opportunities and maybe draw some parallels to how they may differ from agreements that have been signed by some of your peers? I think there's a lot of confusion out there in the market today. And just based on your discussions with some of these counterparties, curious if you anticipate any material changes to economics or demand given the recent headlines?

I'm not surprised that investors are a little bit confused because there's been a lot of numbers thrown out there over the last couple of months. I would think that or it's our position that kind of Core Scientific quarterly deals are probably not likely to be repeated. And especially as you look towards potential hyperscaler customers, deal structures can vary wildly based on how you actually want to position yourselves or how the hyperscaler wants to position themselves. Triple net leases, I think, are pretty common structures and pretty attractive structures from our perspective when we're trying to focus on the infrastructure and the energy aspects. But it's a little early and premature for us to provide any sort of specific as to what those economics would look like other than to say that with the -- with the cost-effective energy and the scale that we have in Pennsylvania as well as the various kind of geographical and infrastructure features of those 3 sites, we have no shortage of interest for those sites. And I think if you look at kind of how the industry has actually performed, you have a lot of sites that have been under development, but you also have a pretty healthy ratio of sites that are being signed up for leasing. The market doesn't seem to be overheated at all on the infrastructure side, at least in our opinion or opinion that we've gotten from a lot of our strategic advisers. It seems to be a very healthy market and robust market. And if you've got a good site and if you're able to develop the infrastructure for it on a timely basis, you're likely going to be able to find a tenant for that site over the next year or 2. That seems to be the general kind of consensus on where the market is. I know the market was a little bit concerned around Microsoft's comments. But really, I think that's one player amongst many. And really, the opportunity here on the HPC and AI infrastructure has not slowed down at all.

For the second question, obviously, we've seen a shift in focus towards the North American portfolio with plans to develop energy and HPC infrastructure. Perhaps you can speak to the remaining portfolio in South America. Is there a potential for further divestment in that region to fund growth in Pennsylvania?

Yes. So these -- the sites that we have in Latin America are up and running. They're profitable sites, generating free cash flow. The site that we had here with Yguazu was a bit of a different situation, right? Because it wasn't up and running. It wasn't generating free cash flow. It didn't already have miners secured for it. And so there was a lot of costs associated with securing the miners and building that site up to the full 200 megawatts. And that's a very different proposition than keeping a profitable and free cash flow generating site online. I think that we don't have an unhealthy emotional attachment to any particular site. I think if the economics were compelling, we're obviously going to look at it and look at all of those opportunities for creating shareholder value that come our way. But it's a bit of a different basket to be looking at an operating free cash flow generating site versus a site that is not operating and have substantial costs associated with bringing it up and running.

Our next question comes from Brian Dobson with Clear Street.

So I mean, in our view, Bitcoin miners have excellent experience for HPC development. But as you're speaking with potential clients, I guess, how do you frame that? And what's the reaction, Ben, about the company's, I guess, past experience?

Yes. It's a good question. I think when you're a Bitcoin miner, you're used to doing deals pretty quickly and pretty aggressively. You're very opportunistic. When you are looking at the HPC space, it's a very different space. People are taking a lot more effort. They take a lot more time, and they do a lot more diligence before they move forward on a deal. And it's very important that they do all of that because these are significantly more capital-intensive projects than Bitcoin mining, right? For 2024, our rough cost to deploy a megawatt of miners and infrastructure was about $1 million a megawatt, whereas for HPC and AI, you're looking at probably anywhere between $30 million and $40 million depending on the GPUs and the servers and other equipment that you're choosing to operate in. And so that level of capital intensity does require a more stringent diligence process. I think if you go into these conversations without having the right resources in hand and without being knowledgeable, it's pretty easy to just not have that go anywhere and for the potential hyperscalers just stop taking your calls. So that's why it's so important for us to engage with the right strategic advisers and recruit the right people internally to help us manage these processes and present a very strong footing forward with potential HPC customers. And so far, it's gone very well with our strategic advisers. And our team, they are filling in the knowledge gaps and really bringing us up to speed very quickly on an educational perspective as to what we need to do, what we need to learn and how we can take our core competencies as a Bitcoin miner and apply them to this new industry with HPC and AI with the greatest efficiency and outcome.

I would now like to turn the call back over to Ben Gagnon for any closing remarks.

Okay. Well, thank you, everyone, for joining today. I'd just like to reiterate one more time just how significant the strategic moves we've made over the last few months are. Across nearly every dimension, we have rapidly transformed this company and are now a U.S.-focused energy and compute company with a strong underlying Bitcoin mining business and exciting potential to expand into HPC and AI in 2025 and beyond. We look forward to keeping you up to date on our progress. Thank you very much.

Thank you. This concludes the conference. Thank you for your participation. You may now disconnect.