Bitfarms - Earnings Call - Q2 2025

Transcript

Speaker 5

Ladies and gentlemen, thank you for standing by. Welcome to Bitfarms' second quarter 2025 earnings conference call. At this time, all participants are in a listen-only mode. After the speaker's presentation, there will be a question and answer session. To ask a question during this session, you will need to press star 11 on your telephone. You will then hear an automated message advising your hand is raised. To withdraw your question, please press star 11 again. Please be advised that today's conference is being recorded. I would now like to turn the conference over to Laine Yonker, Director of Investor Relations. Please go ahead.

Thank you, and welcome to Bitfarms' second quarter 2025 conference call. With me on the call today are Ben Gagnon, Chief Executive Officer and Director, and Jeff Lucas, Chief Financial Officer. Before we begin, please note this call is being webcast with an accompanying slide presentation. Today's press release and our presentation can be accessed on our website, bitfarms.com, under the Investor section. Turning to slide two, 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 the reconciliations to IFRS measures. Please note that all financial references are denominated in US dollars unless otherwise noted. Now, turning to slide three, it is my pleasure to turn the call over to Ben Gagnon, Chief Executive Officer and Director. Ben, please go ahead.

Speaker 4

Good morning, everyone, and welcome to Bitfarms' second quarter 2025 earnings call. We made strong, steady progress in Q2 with several key developments that advanced both our Bitcoin mining and HPC and AI businesses. Today, I'll walk you through these developments and how they position Bitfarms to execute on our HPC and AI growth strategy, through which we believe we'll be able to maximize the value and potential of our energy portfolio. Turning to slide four, I would like to start with an update on the low CapEx Bitcoin mining foundation that is underpinning our transition to an HPC and AI infrastructure company. In the second quarter, we installed more than 12,000 miners, completing all of our Bitcoin mining growth plans and initiatives across all of our facilities. We remined 718 Bitcoin for a direct cost of $48,200 and achieved revenues of $98,000 per Bitcoin.

During the quarter on May 12, we also received notice that our electricity provider in Argentina would be halting service immediately pending negotiations with its creditors. We spent the last three months working through the various options available to profitably resume operations at our Argentina facility. These negotiations are moving slowly with no visible pathway to re-energization. With rising costs in the country following the Argentina IMF deal in April and no electricity supply, the company has made the difficult decision to shut down our mining operation in Argentina by November 11, 2025. Notably, Argentina had our least efficient mining fleet, rising energy costs, and our least reliable electricity supply. While our hash rate is down as a result, the impact on free cash flow is partially offset by the improvement of other fleet-wide key performance indicators, which have improved from the shutdown.

Energy efficiency is up 1%, average electricity price improved 2%, direct hash costs improved 5%, and uptime is up 2%. These improvements help to offset the impact on free cash flow and further de-risk our mining operation through more competitive performance. In early July, we also completed the Bitcoin Miner Repair and Upgrade program, with one final transaction upgrading more than 10,000 T21 miners for approximately 8,500 S21+ miners under the same terms and conditions as previous upgrades. This last round of miner upgrades included all of our new T21 miners from Argentina, as well as T21 miners from other facilities. All of the new S21+ miners were imported into the U.S. in advance of the August 1 tariff rate hike and are actively being marketed for sale.

Our older M50 and S19 miners, as well as various mining equipment from Argentina, have either already been sold or are actively being marketed for sale. We expect estimated proceeds from Argentina's shutdown of approximately $18 million through the elimination of site remediation liabilities, recovery of prepaid deposits, reduction in lease expenses, a free termination option, and equipment sales. This is equivalent to over two years of free cash flow from Argentina mining operations with current economics. This plan is expected to reduce risk and our cost to mine Bitcoin through the improvement of operational efficiencies while improving financial liquidity to support our U.S. HPC and AI growth initiatives in the near term. Turning to slide five, with all mining growth initiatives complete and the Argentina shutdown underway, I would like to provide a snapshot of where we are today.

We are currently operating 17.7 EH/s at $17 per terahash across our 14 data centers, which is generating free cash flow for mining operations of approximately $8 million a month. We continue to sell Bitcoin from our mining operations to fund OpEx, overhead, and CapEx, but we have also increased Bitcoin holdings to approximately 1,200 BTC, up 25% from year-end 2024, representing a value of approximately $145 million with a Bitcoin price of approximately $119,000. Extra hash is pulled back with the Argentina shutdown, our Bitcoin mining business remains a low-risk cash flow foundation with minimal CapEx needs for the foreseeable future and considerable upside to rising Bitcoin prices, enabling us to invest all our time and capital into developing our HPC and AI business. Turning to slide six, I would now like to provide an update of our North American energy portfolio and its HPC and AI potential.

Bitfarms stands out from other public mining companies with a unique energy portfolio built over many years of operation. In North America, we believe we're the largest public miner in Quebec and Canada, have the biggest footprint in Pennsylvania and PJM, and have the largest footprint in central Washington. The geographic diversification across data center hotspots gives us a serious advantage in the HPC and AI race, which is still in the early innings. Here's a portfolio. With over one gigawatt in the pipeline, our Pennsylvania portfolio is comprised of three sites: Panther Creek, Scrubgrass, and Sharon, and is where the bulk of our U.S. portfolio lives.

Amazon has recently committed to investing over $20 billion in their data center campuses less than 40 miles to the northeast and southeast of Panther Creek, and CoreWeave has recently committed to investing over $6 billion in their campus 30 miles to the southwest of Panther Creek, leaving Panther Creek in the center of it all. With robust energy, fiber infrastructure, and recent strong political tailwinds for data centers, which I will speak more to in a second, Pennsylvania is well positioned for massive strategic wins with large hyperscale clients. On the opposite side of the U.S., our Washington portfolio is much smaller at 18 megawatts but is strategically located in the largest data center cluster on the West Coast.

With approximately one gigawatt of data centers in the region, including Microsoft and CyrusOne within a 30-mile radius and some of the lowest cost energy for data centers in the U.S., Washington is well positioned to service prospective HPC and AI enterprise clients and generate attractive margins. North of the border in Quebec, our 170 megawatts of reliable and cost-effective hydropower makes us the largest Bitcoin miner in Quebec and in Canada overall. The province's robust supply of cost-effective renewable electricity and extensive fiber infrastructure to Toronto, the East Coast of the U.S., and Europe has attracted the development of approximately 700 megawatts of traditional data centers in Quebec, representing over 50% of all data centers in Canada. Major players including Microsoft, Google, Amazon, and Vantage are all present and looking to expand in the area.

With recent pushes by federal and provincial governments on data sovereignty and privacy, the development of data centers in Canada has emerged as a rare nonpartisan objective that has strong support through federal and provincial governments. If Bitfarms converted all of our Canadian Bitcoin mining megawatts to HPC, it would represent up to a 24% increase in data center megawatts in the province and make us one of the largest HPC data center operators in Quebec and Canada. In order for us to capitalize on this opportunity, it will require regulatory approval for converting our crypto mining megawatts to traditional data center megawatts. While this has broad political support and we have received initial indications of support at various levels for such a conversion, it will take some time to run its course.

In the meantime, our mining operations in the province continue to be profitable, and we are tentatively planning a conversion of the portfolio in 2027 and 2028 in advance of the next halving event. We will provide further updates on Quebec as they materialize. This unique portfolio is attracting huge interest from counterparties, and our scale and positioning are unmatched. As we replicate the lessons from Panther Creek across our North American sites, we are well positioned to capture meaningful market share in these high-demand HPC markets. Turning to slide seven, in recent weeks, there's been a surge of new data center announcements in Pennsylvania. Big names like Google, Meta, Blackstone, Brookfield, and CoreWeave have committed over $90 billion in investments, citing robust energy and fiber infrastructure and close proximity to the largest data center cluster in the U.S., Data Center Alley, and major East Coast metropolitan centers.

The robust investments into these data centers and the meaningful job growth that accompanies them will be transformational for many towns and counties across the state. As a result, they're receiving robust political support and attention from the local level all the way up to the Oval Office. As the only public Bitcoin miner with a big Pennsylvania footprint, we're perfectly positioned in what is emerging as the new AI hub. This isn't just luck, it's vision. We've been evaluating Pennsylvania for years, focusing on the same opportunity that is currently taking the industry by storm. With surging demand for power, record capacity auctions, and a renewed interest in traditional thermal generating assets, the power plants we acquired in the Stronghold acquisition are worth significantly more now than when we bought them just a few months ago.

Turning to slide eight, our flagship campus in Pennsylvania, Panther Creek, is advancing rapidly. In the past few weeks, the electric utility PPL Corporation has given us the green light for expanding our grid connection, confirming firm service of 50 megawatts by year-end 2026 and an additional 300 megawatts as early as 2027. That means more power on a faster schedule confirmed. To support the development, we've also executed binding purchase and sales agreements that more than double the acreage of the site, paving the way for a contiguous and marquee 350-megawatt HPC campus in eastern Pennsylvania. The conceptual master site plan that you are looking at on screen now is based on the new power schedule and expanded acreage. This master site plan is finished and has been submitted to Macquarie.

It is important to remind everyone that the master site plan may be modified based on the multiple conversations we currently have going with potential customers, but it's a rock-solid base to develop the site and customer conversations expeditiously. Phase one hits 50 megawatts by the end of 2026, and phase two adds 300 megawatts as early as 2027. The plan includes the development of four buildings housing 350 megawatts in total, with room for additional expansion in a phase three with a fifth building once the ongoing FERC, TALEN, and Amazon issues are settled, or we have confirmed an alternative arrangement that permits us to use our existing grid connection and power plants for the HPC data center campus. Finally, our new partnership with T5, a top-tier data center developer who's worked with every major hyperscaler to build and operate data centers, is a big deal.

T5 will be responsible for managing the development of our Panther Creek campus and is yet another proof point to investors and prospective customers that validates the development potential. Next steps for Panther Creek include fast-tracking permits, locking in long-lead items like transformers and generators, and breaking ground as originally planned in Q4 this year. CapEx-wise, the Macquarie facility will be sufficient for phase one, and with just $10 million of planned CapEx for the remainder of this year, the majority of CapEx will fall into 2026. Turning to slide nine, while Washington is much smaller than Pennsylvania, this site is a gem. It's positioned in the West Coast Data Center Alley, and the utility has confirmed that all of our Bitcoin mining megawatts can be converted to HPC with no regulatory red tape.

Better yet, converting our megawatts would cut our energy cost per megawatt-hour by almost 50%, below $30 per megawatt-hour, and make it some of the cheapest power for data centers in the U.S. To help us further advance customer conversations, we have executed a binding PSA for adjacent land that will more than double our acreage and is expected to give us all of the space we need to develop a state-of-the-art HPC facility. In Washington, we're eyeing enterprise customers where the smaller site can be used to greater effect, generating incredibly healthy margins and providing portfolio diversification. Turning to slide 10, following the successful rebalancing of our portfolio earlier this year and the Argentina shutdown, our current operational megawatts are now over 80% North American, and with the majority of our growth pipeline in the U.S., our pivot to the U.S. is in full swing.

Our actions are not limited to the expansion of energy and the development of HPC infrastructure in the U.S. Recently, we've announced the opening of our second Principal Executive Office in New York City and our planned transition to U.S. GAAP accounting for full-year 2025 results. These actions are expected to simplify our reporting processes, reduce administrative and legal costs, broaden our U.S. investor base, and improve our eligibility for inclusion in certain stock indices, among other potential benefits. These are important steps on our journey to becoming a U.S.-domiciled entity, which we plan to achieve in 2026. We strongly believe this strategic transition will better position Bitfarms to execute our HPC and AI growth strategy, driving improved operational efficiencies and maximizing shareholder value. Turning to slide 11, while we have accomplished a lot in the last year, I'm particularly proud of our accomplishments over the last four months.

Our Bitcoin mining business is locked in, projected to produce strong free cash flows of approximately $8 million per month, with minimal CapEx and numerous improvements across key performance indicators. With more than one gigawatt, our Pennsylvania pipeline is a gold mine and one that's quickly emerging as a major AI hub, fueled by massive regional investment. HPC and AI development is taking off with rapid progress at our flagship campus, Panther Creek, backed by top advisors, strategic partners, and financing. We've taken key steps towards a planned 2026 U.S. re-domicile, which will drive index inclusion and expand our access to U.S. investors and capital. Our portfolio has rebalanced to 80% North American, with all U.S.

sites well positioned for conversion to HPC, and we have dramatically strengthened the balance sheet with approximately $230 million in cash and BTC as of August 11, and access to up to an additional $250 million in the Macquarie facility for Panther Creek development. Our confidence in our business and ability to both execute and create long-term value in this transition has never been stronger. Despite our strong performance, we believe that the market is undervaluing both our Bitcoin business and HPC potential. Accordingly, we have recently launched the company's first-ever stock buyback program for up to 49.9 million shares, or about 9% of outstanding shares. I'm proud to say that in the first two weeks, we have already repurchased roughly 5 million shares, equivalent to 10% of the total share buyback program.

With strong cash flows, ample liquidity, and financing, we plan to continue buying back our shares under this program for the foreseeable future. Now, turning to slide 12, I'll turn the call over to Jeff for the financials.

Speaker 2

Thanks, Ben, and good morning, everyone. Turning now to slide 13. Before I dive into the second quarter results, I'd like to highlight the key financial achievements from the quarter. First, the Macquarie financing. At the beginning of the quarter, we entered into a financing agreement with Macquarie to fund the development and build-out of our Panther Creek, Pennsylvania HPC data center project. The facility provides up to $300 million towards the design, development, and initial phase of construction of the project. The initial $15 million tranche was drawn at closing to fund the soft costs associated with the design of the Panther Creek data center, land acquisition, and other initial steps. The $250 million project finance tranche is drawable upon meeting certain conditions, including development milestones.

With just $10 million of planned CapEx for Panther Creek for the remainder of this year, there's no anticipated need to draw down an additional tranche until 2026. We are excited to join forces with Macquarie, an established leader in HPC data center financing. They are truly a partner in our HPC expansion, bringing financial and development expertise to supplement our other partners and advisors. This financing is emblematic of our financing strategy for our total portfolio as we use cost-effective debt to fund our growth. Our highly valued and appreciated North American assets, combined with the higher margin and predictable earning streams characteristics of HPC and AI, make our HPC properties attractive to debt finance. Second, cash generation of our Bitcoin mining operations.

As we've noted previously, during the development and build-out phase of our HPC program, our Bitcoin mining operation funds our G&A and corporate expenses and the debt service requirements associated with our HPC build-out. We are achieving steady mining margins and, as Ben noted, generating about $8 million of free cash flow from operations per month in excess of our corporate administrative costs to fund our initial capital requirements, debt service, as well as our stock buyback program. With our miner upgrade program behind us and having among the most efficient fleet in the business at an average efficiency of 17 watts per terahash, our mining operation is benefiting from the current high Bitcoin price to generate strong cash flows, a solid base for HPC expansion. Third, our stock buyback program launched in the last week of July.

As announced a few weeks ago, our Board of Directors has authorized Bitfarms Ltd. to repurchase up to about 9% of our outstanding shares over the coming year. Our share repurchase program calls for the buyback to be funded by excess cash flow from our mining operations. As Ben noted, we have repurchased over 4.9 million shares to date, and since the inception of the buyback program, our purchases have been at daily volumes that approximate the daily allowable limit for the exchange on which the shares were purchased. We believe we're undervalued versus peers, and we're using mining free cash flow, not our Bitcoin treasury or funds from our debt financing, to prove it.

Whether our shares need to reflect the cash generation potential of our mining activities, nor the value of our HPC properties, utilizing our excess cash flow from operations is an effective way to bring value to our shareholders. I'll add that we have not issued shares under our ATM since January, and the shares we have repurchased to date under our buyback program are well below the prices at which we issued the shares last year. Lastly, our U.S. GAAP transition. As Ben noted, this transition is vitally important as we continue to shift our operations towards the United States. We expect to report our full fiscal year 2025 results on the U.S. GAAP. We believe this will simplify our financial reporting process, further align us with our U.S. HPC and AI and mining peers, and serve as an important stepping stone towards the potential re-domicile in 2026.

The benefits of a re-domicile to the U.S. are plentiful. Bitfarms Ltd. would have access to a broader pool of U.S. capital and investors, potential inclusion in major U.S. stock indices, and more favorable regulatory and tax environments, among other benefits. Building off successful rebalancing of our energy portfolio to over 80% North American megawatts in 2025, these initiatives further advance our U.S. pivot and align with our HPC and AI growth strategy. Turning to slide 14, let's focus now on our second quarter financial performance. In the second quarter, we earned $718 million in Bitcoin and achieved a total revenue of $78 million, up 87% YoY. Revenue from our mining activities was $71 million, and a balance of $7 million was earned from hosting electricity generation businesses following our acquisition of Stronghold Digital Mining and from our Volta Electrical Services subsidiary.

Our gross mining profit was $32 million, representing a direct mining margin of 45% and an average direct cost of $48,200 per Bitcoin mined. Cash general and administrative expense, or G&A, was $18 million, an increase of about $2 million over the first quarter, reflecting Stronghold G&A of approximately $3 million in the second quarter versus about $400,000 for the two weeks in which Bitfarms owned Stronghold in the first quarter. Excluding Stronghold's G&A, our second quarter cash G&A would have been roughly 3% lower than the first quarter. Operating loss was $40 million in the quarter and included $15 million of impairment charges attributable to our Argentina operation. As a result, net loss for the second quarter was $29 million, or $0.05 per share. Turning to slide 15, for the second quarter, our adjusted EBITDA was $14 million, or 18% of revenue. This encompasses primarily our self-mining revenue.

Our direct mining cost for Bitcoin in the second quarter was $48,200, with our all-in cash cost to mine a Bitcoin at $77,100, and paid the revenue per Bitcoin earned during the quarter of $98,000, resulting in a profit per Bitcoin of $20,900, or a profit contribution from mining activities of about $15 million. Turning now to slide 16, as of August 11, we had total liquidity of approximately $230 million comprised of cash and unencumbered BTC. In addition, under the Iguazu sale agreement, we expect to receive the remaining $10 million from Hive by the end of September, marking the completion of the sale announcement earlier this year.

As was mentioned earlier, we project to generate on average about $8 million per month of free cash flow from mining operations, which we expect to be more than sufficient to fund the remainder of the 2025 CapEx, our working capital requirements, our debt service, and the share buyback program. Regarding our Bitcoin One program, we've realized $11 million in profits since the program's inception this past February. This follows the success of our Synthetic Hollow program, under which we realized $18 million in profits between October 2023 and January 2025, with total net profits since the inception of our market operations activities in October 2023 of over $28 million. We are conducting a strategic review of the program as we further shift our focus in operations towards HPC and AI.

In closing, we believe our financial position provides a solid foundation to execute on the HPC and AI initiatives that Ben laid out today, and we look forward to keeping you updated as we continue to develop our U.S. operations. With that, I'll turn the call back to the operator for questions and answers.

Speaker 0

Thank you. As a reminder, to ask a question, please press star 11 on your telephone and wait for your name to be announced. To withdraw your question, please press star 11 again. The first question will come from Nick Giles with B. Riley Securities. Your line is open.

Speaker 2

Thank you very much, operator, and good morning, everyone. This is Cedar Chabalin on behalf of Nick Giles and Ben and team. I just wanted to touch base on Panther Creek. What's the game plan for getting construction procurement lined up? You're planning to break ground in 4Q, and am I right that you're moving ahead regardless of whether you locked in customers at this point? I'm also curious about your Macquarie financing option. What boxes do you need to check to get the debt facility entirely in place? I remember you mentioned that it would be a multi-step process, and if you could ballpark the CapEx you're looking at for the construction phase, that would be really helpful. Thank you very much.

Speaker 4

Thanks, Fedor. I'll speak to the initial questions around the development for Panther Creek, and I'll pass it off to Jeff to answer the questions around Macquarie and CapEx. To answer the question for, you know, what are the next steps for development at Panther Creek, we've recently engaged T5 as owners' reps, so they're managing the process here with engaging all the contractors, overseeing the development, securing the permits, and everything else. There are a few different things that we are going to be doing over the next couple of months, but the biggest one that we'll be doing will be the civil works, beginning to clear out the trees, flattening out the land, as well as the substation construction. Those are the big development milestones prior to the end of the year.

We won't actually start constructing the buildings until the civil works is done, so that'll be in the first part of next year. That's what we should expect to happen over the coming months. Jeff?

Speaker 2

Sure. Good morning, Fedor. In reference to where we stand currently with Panther Creek here, for the phase one of 50 megawatts, our capital expenditures for the balance of this year, 2025, are around $10.5 million, so we're in very good shape for that. For the project in total through 2026, we're envisioning it to be around $400 million in total for the full build-out of the facility. Does that address your question?

Yeah, you cut off a little bit, sorry. I do have a follow-up on there. Keep hearing that hyperscalers are thinking really, they're really big these days, right? They're just looking for gigawatt scale campuses. How does it play into your Panther Creek strategy? Based on the conversations you've been having while marketing the site, what's the typical site profile of the players showing interest here?

Speaker 4

Yeah, it's a good question, Fedor, and you're right. The increase in demand per site is happening quite rapidly. I think last year when Core Scientific and CoreWeave announced their deal, the focus was really on 100 megawatt sites, and that number has gone up quite a bit. I think the demand here is tied also to time. The way that the market is viewing things is that 2026 power is highly valued and prioritized. Because of the timeline on 2026 power, people tend to be a little bit more open to smaller sites in 2026, provided, of course, that there's the ability to expand beyond that over time. Most of the HPC customers that you speak to, they understand that a site is not built all in one go, but it's separated out into multiple phases.

What most customers are looking for is immediate power in 2026, with the ability to expand beyond that over time. What we're seeing is that the larger the campus, the more interesting for sure. You really can't get away from that timing aspect, which is kind of bending a lot of perceptions on what people would like and what they're interested in. With the 350 to 410 megawatts that we have there at Panther Creek, this is still a very sizable data center campus. There's roughly 600 megawatts worth of data centers in the state of Pennsylvania alone today. It would be more than a 50% increase in terms of the total data center capacity in the state just by building out this one campus. Certainly, the demand is tracking upwards.

Speaker 2

Thank you, Ben, Jeff Lucas, and team. Thank you for this and continue a bit of luck. I jump back to you.

Speaker 4

Thank you.

Speaker 0

Our next question will come from Mike Collins with HC Wainwright. Your line is open.

Speaker 3

Hi, good morning, guys. Thank you for taking my questions. Follow up on the CapEx spend on Panther Creek, Jeff. You mentioned the $400 million total build-out cost, which implies $8 million per megawatt. Curious what that entails from a data center readiness perspective and the implications for potential economics that you could secure from a potential deal with this type of build. I know in the past you've talked about powered shell through fully developed and constructed tier three data center, but what does this entail again from the $8 million per megawatt and what you can expect from a potential contract on that site?

Speaker 2

Sure, let me start that off, and then Ben, you can follow up afterwards. I think first of all, that's the number that we're working out at this point in time. That may vary depending on the arrangements that we put in place with every prospective customer that we finally come to an agreement with here. That's actually the expectation we have for a full build-out of having the site ready for larger GPUs here at that point in time. That's the position we have here as we're currently assuming here. That may change again given the nature of whatever arrangement we have in place with our prospective customers. Anything you wish to add?

Speaker 4

Sure. You know, when we're looking at the build-out, we're trying to optimize for two different things here, Mike. You know, we historically have always tried to optimize on return on invested capital, and we're also trying to optimize for the multiples expansion being driven by these kind of long-term contracted revenues that come out of these HPC contracts. I think generally speaking, trying to optimize across those two variables is quite difficult because the best return on invested capital in this industry is going to be for just the development of powered land. You know, if you can just get a piece of land and you can put some 500 megawatts on it, that investment is going to be the best possible return that you can have per dollar invested. The market's not really going to give us value for that transactional nature.

There needs to be a shift further up the stack so that you can get those long-term contracted revenues, which is going to drive that multiples expansion that we're targeting. We're also being, you know, quite realistic and pragmatic here that this is an execution game and that the higher up the stack we go, the more you're going to get from the multiples expansion, the more CapEx you're going to have to raise, and the more difficult it is to execute. By looking, you know, more broadly at what we could possibly do, we have a better way to recognize value out of those portfolio assets through a powered shell or a colo opportunity.

Speaker 3

Very helpful. Thanks for the color, Ben and Jeff. A follow-up for me, as it relates to the potential Bitcoin mining conversions at your Quebec data centers, I know in the future now, but if you could just provide more color in terms of what would need to happen for you to ultimately make that decision, and is there potential to convert the rest of the portfolio if demand is strong enough from prospective HPC AI tenants?

Speaker 4

It's a great question on Quebec. We spent a lot of work and time evaluating the Quebec opportunity. We've had numerous different consultants and strategic partners out there evaluating sites. We think there's a lot of good conversion potential for Quebec. As I mentioned earlier on in the call, we will need to convert over the crypto mining megawatts to data center megawatts, and that requires regulatory approval. There is no clearly defined regulatory approval path because nobody has ever done this particular conversion of megawatts over. The crypto mining megawatt tariff is relatively new, only having been established a few years ago. The huge demand on HPC is also relatively new. Nobody's actually asked Hydro-Quebec to try and convert these prior to us asking Hydro-Quebec to convert these megawatts. I will say that, politically, Quebec has never been a huge fan of Bitcoin or Bitcoin mining.

I think that's largely true for the federal government in Canada as well. They are very interested in HPC, AI, and sovereign data centers. Here's a great opportunity for both Bitfarms to convert over and develop our sites, but also for the relative power brokers and political interests in local, provincial, and federal to all really tick some strategic objectives here with conversion over Bitcoin mining megawatts to data center megawatts. I think everyone's aligned in terms of the value creation opportunity that's here and wanting to get this done. Unfortunately, I can't comment to a specific pathway and a timeline for converting over those megawatts right now. We will provide an update as soon as we have that because we are actively pursuing it.

Speaker 3

Great. Thanks, Ben. That's a lot.

Speaker 4

Thank you.

Speaker 0

The next question will come from Gareth Gasetta with Cantor Fitzgerald. Your line is open.

Speaker 4

Thanks for taking the question. Just a quick one on the Macquarie credit facility. In order to draw down that second $250 million, I guess you have to submit kind of a draft statement to Macquarie. How long do you expect that approval process to take and how long until you're actually able to access the capital?

Speaker 2

The process is going on at the moment here in the review process. While we expect it to take probably a couple of months here, there's always a chance that requirements could push us into early 2026. What I think is very important to point out here, as we've noted here, is that with Panther Creek, we have about $10.5 million of capital expenditure requirements for this year. That and even going to our expenditures into early 2026, we're more than sufficient with our liquidity actually to fund that and then get reimbursed when the other piece comes into play here. So far, things are going well with Macquarie, and they are in the process where they're viewing our master site plan as we presented it to them.

Speaker 4

Got it. Thank you. Just a quick follow-up here. How are you thinking about the levels of share buybacks going forward? Are you thinking kind of consistent to what happened this quarter? Any additional color would be good. Thanks, guys.

Speaker 2

Let me add just a few comments to that as well. You know, we're continuously assessing our cash flow generation for Bitcoin mining activities here. As we've noted, we're generating about $8 million of excess cash flow above our operating requirements, which includes our overall G&A here per month for the balance of the year. That is going to dictate in large part what we actually do with a buyback plan. We are anticipating continuing it in the range of where we are now. Of course, that's going to depend upon the opportunities and the CapEx requirements that arise over the remaining four and a half months of the year.

Speaker 4

Great. Thank you.

Speaker 0

The next question will come from Brian Kinslinger with Alliance Global Partners. Your line is open.

Speaker 2

Great. Thanks so much for taking my question. With the master plan now submitted, at a high level, can you talk about whether potential customers have begun the evaluation process or what has to happen before that begins? How has T5 changed those conversations in general?

Speaker 4

Thanks, Brian. Yeah, we're actively in conversations with potential customers now. This really started in earnest after we had secured the land. The big driver forward now is having the certainty from the ESAs from PPL as to what we can expect from grid connections and when we can expect them. That's been a huge driver forward in these conversations. The other part is having the right counterparty here for helping us to develop out the sites and manage the whole process. T5 Data Centers is a very known data center builder and operator. They work with every major hyperscaler in the space. They're building numerous of these data centers for CoreWeave. This is a trusted party, and that brings a lot of confidence to potential customers when we're having these conversations.

Those two things are really helping to drive this forward: the certainty around the energy supply and delivery date, as well as the increased confidence in the development partner are both key items.

Speaker 2

Great. A follow-up just on the Bitcoin mining business. Is there a way to think about maybe one year out and two years out how you think about what your hash rate under management might be?

Speaker 4

It's a great question for thinking that far out. Really, what we should expect is that we're not planning any mining growth. As we convert over to HPC and AI, there could be some gradual pullbacks in the exahash. In the Q4 deck, one of the things that we outlined was our best expectations for network hash rate and Bitcoin price through every quarter and from 2025 and 2026. I think that is a very good projection for where the economics might be able to go in the future and what that might imply for revenue and gross mining margin for Bitfarms and for the industry at large. If you look at that and you look at the kind of revenues associated with that, we're going to get continued revenue and free cash flow performance out of our mining fleet. At least we expect to throughout the remainder of 2026.

When you look at the relative opportunity between HPC and AI and Bitcoin mining investments right now, it's really hard for us to justify any further investments into Bitcoin mining given the HPC development opportunity and the potential. You can expect that the hash rate should stay flat. As we progress with HPC development, especially at Panther Creek, there may be some gradual reductions in the hash rate over time.

Speaker 2

Great. Thank you.

Speaker 0

Okay. The next question is going to come from Stephen Glagola with Jones Trading. Your line is open.

Speaker 4

Ben and Jeff, thanks for the question. Just following up on an earlier question in Q&A, could you put some concrete figures or guidance around how you think about sort of annual revenue per critical IT megawatt on a powered shell lease at Panther Creek, as well as maybe the corresponding EBITDA margin from this lease structure? Additionally, what, you know, like PUE benchmark would you consider appropriate for this? Thank you. Thanks, Steve. There are a lot of estimates that you can point to for what a revenue figure would be per megawatt per annual, what have you, for these various structures. I don't really want to commit to one right now because it's going to be highly dependent on the customer conversations we have and the structure that gets put in place.

What I think is important to focus on here is that the power that we have is in the right place at the right time, and there's significant value creation opportunity in that and far greater opportunity than any sort of Bitcoin mining economics that we can see or foresee. When it comes to how we're planning out and modeling that, we do have a lot of inputs that come from the customer conversations and our advisors, but it's a little premature to be estimating and forecasting things like EBITDA without those customer conversations firmed into a term sheet and a contract.

Speaker 3

Thanks. If I could just ask one more follow-up on the T5 Data Centers, I know they're overseeing all the contracting, permitting, construction. Maybe just elaborate on the cost implications of this. I think specifically, I believe you guys are getting a flat fee. Maybe any additional detail on how that fee is structured and color around how CapEx at Panther Creek is impacted by this arrangement. Thank you.

Speaker 4

Thanks, Steve. We don't expect this to be materially impacting CapEx beyond the estimates that we provided earlier. The estimates that Jeff had provided on the $400 million, that's inclusive of those fees. Really, the key things here with time to market is execution. Bringing in the right parties who know how to execute, who can deliver on time, we should expect to actually pay for itself in terms of higher performance, higher delivery, higher certainty, and less cost overruns from being behind schedule and over budget.

Speaker 3

Thanks, Ben.

Speaker 2

By the way, Steve, one comment you asked about the PUE. Generally, the PUEs that you see in our industry are around 1.5 in that range. By virtue of where our operations and our properties are located here, we're sort of looking towards and projecting a PUE of around 1.25. It's a little better than the industry overall.

Speaker 3

Great. Thanks, Jeff.

Speaker 0

Our next question will come from Mike Crondo with Northland. Your line is open.

Speaker 1

Hey, guys. Thanks. A couple questions. The first one, Ben, you talked a little bit about it on the last question, but why did you select T5? Is there a specific economic arrangement you can share there? Maybe secondly, I'll just ask it now. Related to Panther Creek, did you guys disclose how much the land, that binding agreement to purchase the land, how much that was, and when you would expect that to close?

Speaker 4

Sure. Thanks, Mike, for the question. Just answering the first question here on the land, I believe it's 178 acres. Total value here is, I think, $3.5 million. We should expect it to close this quarter or shortly after. This is happening. We've done all of our due diligence on this site. Everything is basically just the formal closing process now on the land and should be done in advance of us beginning our construction and beginning all those civil works in the coming weeks. Sorry, Mike, I'm blanking on the first part of your question. Could you repeat it?

Speaker 1

Oh, just why T5 and the economic arrangement there?

Speaker 4

Of course. We ran through a process here. We did a bidding process. T5 Data Centers was one of, I think, four different companies who were bidding into the opportunity. We did site visits to multiple different data centers built by all of these different data center companies. We really liked T5's build quality. We really liked their team. They've got a fantastic reputation in the industry. They've got great experience working with every single major hyperscaler. For the kind of customers that we're targeting at the Panther Creek site, T5 knows exactly what the customer wants, how they build, and they're a trusted party to do it. During the process for selecting and bidding, they came in at a very competitive price with this very competitive offering.

The other thing that I'd like to add on T5 is that they're one of the few companies out there that do a full-service offering. This is everything from managing the design, engineering, and construction all the way through to the operations and maintenance. This is one partner that, as we progress with them and as we get more comfortable and more confident in their abilities, there may be opportunities to extend that relationship with them throughout the process longer so that they can provide more value and more services as we do the build-out.

Speaker 1

Got it. One for Jeff. Jeff, the $18 million that I think comes back to you from Paraguay, and it looks like you have some miners, roughly $25 million for sale. When does Bitfarms get that $18 million and the roughly $25 million, when would you expect to have that?

Speaker 4

Okay. Jeff, do you want me to take this one? Because it's one of the—

Speaker 3

Go ahead.

Speaker 4

Go ahead.

Speaker 3

We have the majority of that $18 million comes from miner sales. All of those miners have been imported into the U.S. and are in the warehouse being actively marketed for sale. You know, not entirely sure exactly when those miners would sell, but they're brand new miners, and brand new miners are generally pretty liquid, and they sell for fair market value. We did that upgrade process in order to make sure that when we were going through the sale process, we're selling brand new top-of-the-line miners and not, you know, used miners at a discount. I think we can expect the revenues from that to be coming in over the next two to three months, which could be a little bit faster, it could be a little bit longer, but I think that's a rough timeline for that.

With regards to the remaining funds, the majority of the remaining of that $18 million comes from our prepayment in electricity. We had about $3.5 million in prepaid electricity at the Argentina site, which is now going to be recovered through the latest agreement, and that's going to be paid out over the span of 18 months in a straight line.

Speaker 1

Got it. Just one for clarity, the $18 million from Paraguay and the $25 million you mentioned with miner sales, are those two separate numbers, or am I kind of double counting them?

Speaker 4

To clarify, it's $18 million from Argentina and not from Paraguay. There is some double counting in those figures. The rough expectation on miners from Argentina specifically is about $13 million out of the $25 million total miners available for sale.

Speaker 1

Okay, thank you.

Speaker 0

Our next question will come from Martin Toner with ATB Capital Markets. Your line is open.

Speaker 3

Good morning. Thanks for taking my question. Is there anything preventing you from doing a deal on Panther Creek right now? On how many megawatts would you say you could potentially do, based on that, based on whatever timeline you give me? Same question for Washington. Thanks.

Speaker 4

Hey, there's nothing preventing us from doing a deal now. We're actively in conversations with different customers. It's just that these deals take a lot longer than a Bitcoin mining transaction, which can be done in the span of a week or two. With Bitcoin mining, it's just really as simple as negotiating with Bitmain or MicroBT. When we're talking with the kind of customers that we're talking with at Panther Creek, these are very large organizations with processes in place and a very strong system of procedures and rules and due diligence that they need to follow. There's nothing preventing us from doing a deal. Now we have the power certainty, we have the land, we have the plans in place. Really, it's dependent on the customer's willingness to come in and the customer's appetite for power and time. The same thing is true for Washington.

The difference with Washington is that the site is quite a bit smaller than Panther Creek because we just have just shy of 20 megawatts there. This is still a really, really attractive site. When you look at where the data centers are clustered in the United States, this is the largest data center cluster on the West Coast because it services Vancouver, Seattle, and Portland. It has great proximity to basically every major fiber optic cable, which carries off data to Asia. There is a robust amount of hydroelectric power in the region. This is some of the cheapest power in the United States for data centers, and it's all renewable. It drives a huge amount of demand.

There, we'd be looking at targeting an enterprise customer, somebody who's looking to expand their compute capacity and not really a data center company, but they're using the data centers to facilitate a business or a service or an analytical purpose for their company. We're actively in those conversations with potential customers right now as well for Washington. It's going to be a slightly different structure for Washington due to the relatively smaller scale and the different type of customer that we're looking to engage with. There is no barrier to developing a contract or a deal now. We have everything in place to do that.

Speaker 3

Thanks, Ben. What are the prospects of adding to the portfolio in the Northwest? Is that something that you guys are actively looking at?

Speaker 4

There's plenty of opportunities to continue to grow our energy pipeline. We have a fairly active corp debt team. I think our biggest focus right now is twofold. It's one, it's on executing on the power that we already have. We have over a gigawatt in the pipeline of highly valuable megawatts in the right locations, and that is the big opportunity that we're trying to execute on right now. The priority is execution first, growth second. On the growth side, where we're looking to focus most of our efforts really is on developing the power story at Scrubgrass. We didn't speak to it much on the call, but we believe that there is a big opportunity to expand Scrubgrass beyond what we've already announced. We don't have a firm certainty as to exactly how much that can be expanded, but we believe it can go well beyond 500 megawatts.

Our focus on the growth side is really developing the energy story at Scrubgrass right now, but we're also evaluating other opportunities in the area.

Speaker 3

Super. Last one for me. Is there a time in 2027 when you envision being able to get your first dollar of HPC revenue at Panther Creek?

Speaker 4

It's really going to be dependent on those customer conversations, but we will have power energized by the back half of 2026 at Panther Creek, and we expect to have the first building up by the end of 2026 as well. It's really going to be dependent on those customer conversations that we're having right now as to exactly when that first revenue is going to come in.

Speaker 3

Great. Thank you for taking the questions. That's it for me.

Speaker 0

Our next question will come from Brian Dobson with Clear Street. Your line is open.

Speaker 3

Good morning. You repurchased shares in the quarter. Certainly, there's a lot of pressure on the stock related to Riot selling down its position. Are you looking to take advantage of that expected selling pressure going forward via your repo?

Speaker 4

We believe our shares are undervalued for a couple of reasons, and that's why we've announced the share buyback program, and we're already executing on it. The opportunity is here today. We're not entirely sure how long we should expect to be undervalued by the market, but I think when our shares are so clearly undervalued relative to competitors and relative to the HPC opportunity, it's a really easy call to make to continue buying back shares in the open market. It's a great use of capital, and it's a great thing for shareholders and for shareholder value. It's something that we believe really strongly in. We believe in this company, and we believe in the potential and this management team. Until the market changes its view, we should be active in the market buying back shares.

Speaker 3

Yeah, thanks very much for that color. Just as a second question, there's some very exciting tailwinds in Pennsylvania. Can you maybe elaborate a little bit about how your conversations with potential clients have changed or improved since those tailwinds were announced by the administration?

Speaker 4

Yeah, Pennsylvania is really emerging as the AI hub here, I think, for several reasons. You know, it's a huge energy state. It has great proximity to Data Center Alley in Virginia, DC, New York City, Philly, all these major markets. It's in really close proximity to, with robust energy and fiber infrastructure. It also has pretty close proximity to all the undersea fiber optic cables that carry data off to Europe and Africa. The interest in Pennsylvania is not really surprising to us given how long we've been looking at it. It's a great underwriting of the development potential when we see the major players like Amazon and CoreWeave really putting serious capital down in all of these areas surrounding our site. Amazon has just committed $20 billion to two sites both within 40 miles, CoreWeave $6 billion to a site less than 30 miles away.

I think that's helping to underwrite the value and the development potential here because this isn't just, you know, Bitfarms saying the potential is here. This is everyone saying the potential is here and putting their money down in a very major way. I think it's definitely helping with customer conversations. It always helps to have another counterparty kind of underwrite the same value opportunity.

Speaker 3

Great. Thank you very much.

Speaker 0

The next question will come from Bill Papanasia with KBW. Your line is open.

Speaker 3

Good morning. Thank you for taking my questions. Perhaps you could just walk us through the Quebec portfolio and make note of, you know, perhaps your top three or five sites that you think are most attractive to potentially convert into AI HPC. Thank you.

Speaker 4

Yeah, it's a great question, Bill. The biggest area for us to convert in Quebec right now is the town of Sherbrooke. That's the one that we're prioritizing the most. You know, we have 170 megawatts in the province. Between the three sites that we have at Sherbrooke, there's 96 megawatts there. That has the highest amount of concentration for any of our sites, and we also have the most amount of scale. Sherbrooke is a really interesting city. It's one of the largest cities outside of Montreal in what's called kind of the Eastern Townships area, and it's backed by a huge college campus and technical school there. There is robust fiber and power infrastructure connected with the school and connected with the city.

That is really going to be the lowest hanging fruit and highest value opportunity that we're focusing on developing in Quebec, those 96 megawatts that we have in Sherbrooke. I think that's where we have the most political support from the local community. That's where we have the most economic incentive from the local community, and it's also where we have the best opportunity to deploy at scale.

Speaker 3

Great, thank you.

Speaker 0

I show no further questions at this time. I would now like to turn the call back over to Ben for closing remarks.

Speaker 4

Thank you, everyone, for joining today. I'd just like to quickly reiterate our team's huge excitement about the position that we've been building. We're quickly becoming a U.S.-focused energy and compute infrastructure company, and we have a really strong underlying Bitcoin mining business with a lot of exciting potential to HPC and AI. We look forward to keeping you up to date on our progress as we continue, and thank you very much again for joining the call.

Speaker 0

This does conclude today's conference call, and thank you for participating. You may now disconnect.