Marathon Digital Holdings, Inc. (MARA) Q1 2025 Earnings Report, Transcript and Summary

Good day, ladies and gentlemen, and welcome to MARA's Holdings Quarter 1 2025 Earnings Call. All lines have been placed on listen-only mode. [Operator Instructions] At this time, it is my pleasure to turn the floor over to your host, Robert Samuels, VP of Investor Relations. Sir, the floor is yours.

Thank you, operator. Good afternoon, and welcome to MARA's first quarter 2025 earnings call. Thank you for joining us today. With me on today's call are our Chairman and Chief Executive Officer, Fred Thiel, and our Chief Financial Officer, Salman Khan. Today's call includes forward-looking statements, including those about our growth plans, liquidity and financial performance. These involve risks and uncertainties, and actual results may differ materially. We disclaim any obligation to update these statements, except as required by law. For more details, see the Risk Factors section of our latest 10-K and other SEC filings. We'll also reference non-GAAP financial measures like adjusted EBITDA and return on capital employed, which we believe are important indicators of MARA's operating performance because they exclude certain items that we do not believe directly reflect our core operations. Please see our earnings release for reconciliations to the most comparable GAAP measures. We hope you've had the chance to read our Shareholder Letter and look forward to your feedback. We'll begin with some prepared remarks from Fred and Salman, after their comments, we are going to be conducting an analyst interview with management. Today's session will be conducted by Stephen Glagola and analyst at Jones Trading. With that out of the way, I'm going to turn the call over to Fred to kick things off. Fred?

Thank you. Good afternoon, everyone and thanks for joining us. Despite a volatile start to the year, with global markets pulling back much of the momentum they gained ahead of President Trump's inauguration, Bitcoin has shown remarkable resilience. At MARA, we believe Bitcoin is increasingly positioned as a macro hedge, much like gold offering protection in uncertain environments. Our stock tends to move with the overall mood of the market, especially when it comes to how investors feel about Bitcoin. In risk-off environments, our stock often trades down with the rest of the sector, even when our fundamentals remain solid. This correlation is largely tied to Bitcoin's price and broad sentiment towards miners. We believe the market is fully recognizing the strength of our core mining business, one of the largest in the sector. Meanwhile, some of our peers with smaller operations and little to no Bitcoin holdings are being assigned greater value, which doesn't reflect the full picture. Some of our competitors have publicly announced plans to limit mining growth. or pivot towards hosting and AI GPU operations, while others are expanding by staying greatly attached, investing in new machines, making them less capital efficient. Meanwhile, we are focused on delivering long-term, low-cost energy solutions that outlast market cycles. We're doing this by transforming MARA into a vertically integrated digital energy and infrastructure company. This model gives us tighter operational control, improved cost effectiveness and strengthens our vision -- sorry, strengthens our resilience against broader economic shifts. To achieve this transformation, we're focused on two key priorities: strategic growth with low-cost energy and efficient capital deployment and advancing research and development of digital energy technologies. In 2021, we predicted a paradigm shift in our industry where power producers and miners would converge to utilize an excess energy, especially during periods of low demand, thereby stabilizing the grid and maximizing profits. We've now seen that vision begin to take hold with the power industry acknowledging the benefits of Bitcoin mining as a flexible low. To deliver on this vision, we are in discussions with governments and global energy corporations in the U.S. Europe and the Middle East, which could result in substantially lower energy costs and greater CapEx efficiency. Chasing exahash targets at grid attached prices invites short-term thinking, and we're focused on long-term partnerships to achieve and maintain a low-cost energy strategy. In Q1, MARA delivered strong operational metrics. Our purchased energy cost per Bitcoin was $35,728. Our daily cost per petahash improved 25% year-over-year. And at the start of this quarter, we completed construction at our fully-owned 200-megawatt data center in Ohio with 100 megawatts now online and over 12,000 S21 Pro miners installed. In April, we fully energized 25 megawatts of gas to power operations in North Dakota and Texas, monetizing excess gas and mitigating methane emissions for the producers. And we are now in the process of building out our 114-megawatt wind farm site in Texas, giving us behind the meter low-cost power to extend the life of legacy hardware that otherwise would have been retired. We plan to be fully operational at that site in the second half of this year. Through owning infrastructure, self-generating power, extending asset longevity and managing costs, we are driving capital and operational proficiency across our business. Regarding our second key priority, MARA is investing in and developing digital energy technologies, which can both improve the efficiency of our operations and diversify our revenue streams. As many of you are aware, we're a founding investor in Auradine a U.S.-based ASIC manufacturer whose chips are more power efficient than current market offerings. We have now begun manufacturing custom miners specific to MARA’s needs and applications, providing unique cost and performance advantages. Our hardware 2PIC, 2 phase in merchant cooling is now being deployed at two customer sites, one domestic and one international with a 30-megawatt pilot underway. And preliminary data indicates that 2PIC is successful in enabling miners to increase operational hash rates with minimal losses in productivity -- potentially reducing CapEx by up to one-third for typical mining installation. Our software MARA Pool enables us to capture higher revenues and keep 100% of transaction fees with Block rewards outperforming the network average by over 10%. Looking ahead, we're positioning MARA solutions to support AI inference workloads, which require modular low latency infrastructure. We're already in advanced talks with several compute OEMs to roll out pilots this year. 2025 is an important year for MARA as we continue our transformation into a vertically integrated digital asset digital energy and infrastructure company, and we look forward to sharing more news as the year progresses. In conclusion, in Q1, we reduced costs, increased profitability and proved we can design, build and operate infrastructure powered by self-owned energy. Thank you for your support as we continue our transformation. We've come a long way, but this is only the beginning. Now let me turn the call over to Salman for some highlights from the quarter. Salman?

Thank you, Fred. Q1 was a challenging quarter as the average price of Bitcoin declined meaningfully compared to Q4 and greater network difficulty combined with rising global hash rate impacted our production. Despite these recent headwinds, over the past year, our energized hash rate increased 95% to 54.3 exahash per second in Q1 of 2025 from 27.8 exahash per second in Q1 of 2024. Also, we currently hold over 48,000 bitcoin on our balance sheet. Since our full HODL announcement in July of last year, the price of Bitcoin has increased approximately 52%. At MARA's [Quint Turbo] (ph) strategy of accumulating Bitcoin through mining and purchases, we expect it to create significant value for our shareholders over time. By mining Bitcoin at one of the largest -- I'm sorry, by mining Bitcoin at one of the lowest direct energy cost per coin in the sector, while also opportunistically acquiring Bitcoin, we can capture upside from both operational profitability and Bitcoin price appreciation. Now let me provide some financial highlights for the quarter. I would first like to point out that since the beginning of 2024, we have transformed MARA from 0% owned and operated capacity to approximately 70%. As a result, we made changes to how we present our financial results to give investors a clearer view of our business. The goal with these changes is to make our financials more transparent and better reflect the way our business is evolving. We also reclassified our prior period numbers to be consistent with the new format. These updates didn't affect our reported financial position, results or cash flows. Revenues increased 30% to $213.9 million from $165.2 million in the first quarter of 2024. Average price of Bitcoin was 77% higher this quarter than the prior period -- prior year period and continued -- and contributed about $90.7 million to our revenues, which was partially offset by a decrease in Bitcoin production due to having an increased network difficulty. On the other hand, Bitcoin price as of March 31, 2025, was 12% lower than as of December 31, 2024. This resulted in an unrealized fair market value loss of $510.2 million in Q1 of 2025. As a result, we reported a net loss of $533.4 million or negative $1.55 per diluted share in the quarter compared to a net income of $337.2 million or $1.26 per diluted share in the first quarter of last year. Importantly, since the quarter -- since the end of the quarter, we've seen a substantial recovery in the price of Bitcoin. If that strength continues or even hold steady, we expect it to have a meaningful positive impact on our second quarter results. Despite recognizing a loss in Q1 due to a quarter end Bitcoin price of roughly $82,000 per coin the current approximately over $100,000 per coin price would imply a fair value gain of over $800 million approximately thus far in Q2. Our adjusted EBITDA decreased to a loss of $483.6 million in Q1 2025 from $542.1 million in Q1 of 2024. Finally, our purchase energy cost per Bitcoin for owned mining sites was $35,728 per coin and $0.04 per kilowatt hour. Cost of revenue per petahash per day, excluding depreciation, declined 10% sequentially. Despite the higher difficulty level to mine due to a higher global hash rate. Due to our shift from an asset-light to a vertically integrated strategy, we believe we are well-positioned to reduce our operating costs over time as we further expand our own initiatives. As Fred mentioned earlier, we now operate 139 megawatts of stranded or underutilized energy generation. Since the beginning of 2025, we have completed the acquisition of 114-megawatt wind farm and fully energized our 25-megawatt on-site and edge compute operations at oil and natural gas fields in Texas and North Dakota. These low cost projects are expected to further reduce our mining costs over time. With that, I'll turn it over to Stephen Glagola from Jones Trading to start our management interview. Stephen?

Thank you, Fred, Salman, and Rob for having me here today. I appreciate it. Fred, you opened your shareholder letter and prepared remarks opining on Bitcoin price action and how MARA's stock tends to trade more off-market sentiment versus fundamentals. Bitcoin is up around, I think, 9% year-to-date or so hash price is down 4%, but MARA is down 15%. And so maybe can you just elaborate on where you see the market disconnect between your fundamentals and share performance and why you see Bitcoin price behaving less like a risk asset going forward.

Sorry, let's take that in two parts. So just relative to the share price performance, it's interesting. It seems that the market values us for our Bitcoin holdings but gives us no credit for our Bitcoin mining operations. And if you were to look at how some of our peers who don't have a HODL are valued -- they certainly have a market valuation for the Bitcoin mining operations. So we should, in theory, get the benefit of both, but then that's just how the market happens to see as they look at our big HODL and they value us that way. We personally think that the company is obviously more valuable than that. As regards to the Bitcoin price action, if you would, many people I've seen in the news lately. There's now four or five new companies that have announced that they intend to become Bitcoin treasury companies. You've seen that the state of new ventures now approved their Bitcoin strategic reserve, and they will be actually investing in Bitcoin and buying it. The state of Arizona is going to allocate Bitcoin -- to the Bitcoin strategic reserve that they're creating that is in the possession of state government law enforcement and others. And I fully expect to see other states continue in this path of establishing strategic Bitcoin reserves. In the very near future here, we'll see the administration in Washington announced how much Bitcoin is going into the strategic reserve because I believe we're coming up on the deadline for that. And I think we are starting to see continued good institutional inflows into the ETFs, and we're now seeing a number that was published earlier this week or last week that showed about 9% of existing Bitcoin is now held by corporates and ETFs. And so as that continues to happen, that is going to continue to drive the price of Bitcoin up. And so I believe that we are in a very different market regime than previous regimes where we had cycles that every four years Bitcoin would act based on halvings and miner production. But if you look at the market today, even though most miners are selling and MARA is still doing a full HODL strategy. The price of Bitcoin is moving up. So there's clearly demand in excess of supply. And if you think of stock-to-flow ratios, 450 new Bitcoin being emitted today, if every miner, we're selling Bitcoin that would be 400 Bitcoins going into the market every day and yet a greater number than that are being purchased every day by just a handful of players. So I think we're going to continue to see things move up in Bitcoin longer term, which is why we are still in very much a full HODL position.

Thank you. And I'd like to focus, I guess, my lead questions here on your off-grade expansion strategy. So first, on the types of off-grid sites, MARA's prioritizing for future growth. Can you elaborate on the public private partnerships with governments and energy companies that you're currently in discussions with? And -- are you focusing more on flare gas or wind or other sources for new acquisitions and developments? Thanks.

So if you think about what I said four years ago, I think it was at the Mining Disrupt Conference in July of 2021. I said that Bitcoin miners are going to have to partner with energy companies or energy companies or going to own them. And that is because I believe that eventually grid attached Bitcoin mining, where you have a PPA and you are paying an average cost, which I believe today the average PPA for a Bitcoin miner is somewhere north of $0.04 a kilowatt hour would not be tenable as you start getting out into the years of 2028 and having in 2032, and that you would have to partner with energy companies and find a different way to structure the business. So without divulging details, we are very much executing on that vision of partnering with energy companies. If you look at the Duke University study that was recently done, they published statistic basically saying that while the AI-industry needs somewhere in excess of 40 gigawatts of power, to deploy their growth plans. The U.S. energy grid has in excess of 70 gigawatts available if the load is willing to be flexible by a small percentage, which currently the AI industry is not. We believe that by bringing solutions for load balancing to the AI industry that will enable AI data centers to co-locate with Bitcoin miners, where Bitcoin miners provide the flexible load that enables AI data centers to operate as a flexible load on their grid. This will opened up huge amounts of power to the AI industry. And as we start shifting to inference or what is now more being called agentic AI, those types of data centers don't need 500, 600, 700 megawatts. And they need less than 5. They need less than 50. These are much smaller data centers located in close proximity to the customers that are operating them, primarily private cloud operations, near-prime or on-prime. So we believe that the future of energy is about partnering with energy companies, where we can help them monetize every stranded electron or every underutilized electron. And if you look in markets like the Middle East, for example, when we started working in UAE in the winter time, they had upwards of 3 gigawatts of excess power available and in the summertime, that dropped fairly dramatically. But since then, they have continued to add energy generating capacity. If you look at Saudi Arabia, they likely have 4 to 5 gigawatts of excess energy capacity today that could be made available and monetized and used. If you look in Europe, you have a similar situation. While people reads and headlines about Spain and Portugal going offline due to grid instability issues. They have built out huge amounts of renewable energy, which has not really selling energy during a great portion of the year. And those are all electrons that are going underutilized. And so our focus is to work together with energy companies work together with sovereigns and government entities to take advantage of all of those underutilized electrons by structuring business relationships that enable us to partner with them on a long-term basis to build, operate and manage essentially flexible loads that can help monetize that stranded energy. As you look at the mix, what you're going to find is a mix of thermal, a mix of wind, solar and some flare gas. It really depends on the market and the partner. We're in discussions with some of the largest energy companies in the world that have a mix of all those energy sources and nuclear. And their concern is one of how do they balance load owner grid how do they deploy flexible baseload and how do they monetize from behind the meter at renewable sites, the electrons they can't sell today. And so you'll continue to see a broad mix of energy-generating assets. In regards to flare gas, there are a lot of gas assets around the world that are very applicable to this method. And what I think you'll see us doing more and more in the future is as we continue to work with especially oil and gas producers, you'll see chunks of this flare gas type generation come online in different parts of the world where we're able to deploy our Bitcoin mining operations, as a way to monetize that stranded gas. And we are super excited about those opportunities. But I think the message here is expect us to grow aggressively in partnership with these public and private entities across the world. And as you may recall, one of our goals -- stated goals was to grow our international business such that we had about a 50-50 mix between U.S. and international and we'll continue to do that. And we believe that is a great way of delivering on growing U.S. tax rates because even though we operate outside the U.S., it doesn't mean we won't be pointing that hash rate back to U.S. controlled pools. So we're very excited about being able to contribute to U.S. global mining share growth, across all the countries that we’re mining.

Thanks Fred. Any guidance on the time-line of when we should expect those partnerships to start coming to fruition? Like is that over the course of 2025 starting? Or is that I know it's a 3-year target that you're looking at, but just any context you could provide there?

We purposely not wanted to guide to exahash numbers because what that does is causes the market to look at us chasing a specific number and delivering on a certain time. And when you're working with these large government and energy majors, things can take longer than one thing. I think what I can say, though, is that you will see good progress this year and some announcements that will give an indication of the potential of what's going to happen over the next few years. And I think you'll see us showing up at events in places where other miners normally wouldn't be as we talk to these partners.

Thank you. Maybe this one for Salman or Fred whoever wants to take it. MARA reported a hash cost of [$0.0285] (ph) per terahash in Q1, and that's I think it marks four consecutive quarters of sequential improvement. As you continue to expand this off-grade power capacity, how do you expect this to impact your hash cost position over the next 3 years?

Salman, do you want to take that?

Yes, so Stephen, as you mentioned, we have sequentially reduced our hash cost and our eyes have been fixated on capital and operational efficiencies. And that result is showing up in our historical P&L. As we have stated in the past, we expect our hash cost to continue to decline over a period of time. And those are as a result of actionable projects that we have worked on some of those that you have seen publicly announced and others, as Fred mentioned, are in the pipeline. As we go -- as we move forward from here, we expect the Bitcoin mining to be -- this sector, every four years, there's a having event that happens. And the expectation is the bitcoin price is going to go up. We are building our company in a position where we don't have to rely on the Bitcoin price or hash price to continue to go up. We expect it to go up, and we expect it to create significant value for us, but we are prepared for the downside situation. And that's why it's very important for us to be sourcing megawatts that are closer to the low-cost environment. For example, the two announcements that we had in Q1, the 139 megawatts generation of electricity from our own source. One of them is the wind farm in Texas and that's about 114-megawatt plant, and that results in a significant reduction in our cost because we consume electricity during the time when the wind blows and then the grid is congested. And the grid loves that because we are able to decongest the grid, while hashing at the peak hours, and we don't have to mine when the electricity is expensive. So that's the kind of model that helps us drive our costs down. One point that I want to mention is that while the sector is focused on specific KPIs, for example, some companies, some of our peers are very focused on joules per Terahash. Some of them are very focused on uptime. No doubt that these are important metrics but you have to view what does -- what value does that create for your stockholders? And our focus is the value creation for our stockholders. The uptime could be compromised for that the joules per Terahash could be lower depending on the pricing where you get there, as long as it creates the most value for our stockholders, and that's where our eyes are focused on, and that's exactly what we believe is going to drive our hash cost further down from here.

Great. Thank you. And -- you can think of it as when you have your own generation and your energy price is substantially lower than $0.03 or $0.02 a kilowatt hour then all of a sudden, it opens up the ability to use maybe slightly less efficient machines as some onset. And if we're able to deprecate machines that have already been fully depreciated from our grid attach sites and reuse those behind the miner at these intermittent sites -- your hash cost drops very significantly. And if you don't have the need to operate 99% of uptime because the machines are depreciated, and so they've already been paid for. And so every kilowatt hour you can operate, especially that you can operate at these low power costs is a profitable hour. And so we believe that over time, as we continue to grow this fleet of what we call our AARP fleet, our advanced ASIC replacement fleet or retirement fleet, then that will have a very positive impact on the overall hash cost going forward. It also gives us a secondary advantage, which is if you're only doing grid attached like some of our peers do, every 3 or 4 years, you have to replace your fleet to stay competitive because you're paying a fixed price PPA and potentially even at energy cost is going up due to demand on the grid. And then you sell off those machines and you buy new machines, by being able to retire our machines into a graceful program of retirement at these behind the [meter] (ph) sites, our hash rate hasn't decreased in the same way that our peers is, where they're having to do net zero replacements. We're able to maintain a certain amount of hash rate and by incremental hash rate to add that. It also gives us the ability to even buy other people's retired machines. And if you can buy a machine that is only one-third less efficient than a state-of-the-art machine, but you're only going to pay one-third of the price for it that's actually very beneficial. So we believe this is the right strategy going forward as we continue to scale.

Thank you. And can you address sort of the concerns that lower hash cost for the off-grid mining may be offset by trade-offs like reduced you kind of touched on this, reduced rig uptime, higher operational expenses and elevated capital expenditures sort of maybe how do you see the IRR for off-grid Bitcoin mining sites compared to grid connected sites. And I know you touched on some of the economics of the wind farm in the shareholder letter. So maybe just elaborating on that.

Yes. I think if you look at -- in the case of the wind farm, for example the infrastructure cost as stated in the shareholder letter, I think about $150,000 to $160,000 per megawatt, which is substantially less than what you would be paying in a traditional grid attached because we're able to use these let's just call it retired infrastructure, meaning miners and containers, et cetera. And when you marry that with -- at a wind farm, you don't have an input cost for your energy right, at a wind farm, your energy cost marginally is just to put a number out there, say, $0.01 per kilowatt hour for your O&M. That essentially gives you best-in-class operating cost when you look at the minimal cost of the infrastructure and the power cost. So if you can operate 60% of the time, that's a great opportunity right there. If you can operate 70% of the time, it's amazing. And some of these wind assets have fairly good uptime which allows you to operate that way. In the case of the flare gas, you're operating 24/7 at very low cost. And so those types of sites are, again, very profitable. And as you go to these public-private partnerships, you can operate intermittently, especially if the price of energy that you're paying is similar to that of a wind farm or a flare gas side. So we believe that this concept of Bitcoin going from being a 99% uptime operation to becoming more of an intermittent operation that's able to shape load to meet the demands of the energy companies, you become a very attractive partner for the energy companies, but it also requires that you have systems that allow you to shape your load to the energy companies supply. And that's been a big focus of our technology team over the past year.

Just to add to that, the ARP program that we talked about earlier, the IRRs for those are significantly higher than a traditional model. And even with the depressed hash price assumption, we expect the IRR to be in the range of 30% to 40%. And this is the depressed price environment on the hash price, so very conservative. When it comes to the cost per coin or energy cost you think about, as Fred mentioned electricity costs about $40 per megawatt hour -- all-in cost, probably $50, $55 per megawatt hour for traditional Bitcoin mining out there today. For the ARP program, we expect that to be approximately all-in cost, operating costs around $10 per megawatt hour. So as you can see, were systematically going into the direction of going closer to the lower cost operations that will drive our costs further down from here as we had talked about previously.

Thanks, Salman. I want to turn to AI real quick in the 2PIC. I mean, can you maybe share more early market feedback from your 30-megawatt 2PIC immersion pilot project and what's going on there? And then what specific revenue opportunities do you see emerging from commercial deployments of 2PIC and AI data centers?

Sure. So the 2PIC deployments we've done to date have been done with Bitcoin mining operations. And that's kind of an ideal way to wring out the any bugs, if you would, of these systems. And what we're finding there is that we have an ability to substantially overclock systems. So if you think of it this way, if you need 100-miners operating at nameplate capacity, to essentially operate a site and you have the ability to overclock them by 50%, it means you could reduce the number of miners you actually have to buy by third and miners being one of the most expensive components in a deployment, that CapEx savings is substantial, and the savings is significantly outweighs any marginal increase in cost of the 2-phase immersion costs. In AI, it's even more that way because of your ability to overclock your GPUs as you start moving to 1 kilowatt chips in AI, you pass to move to new types of cooling technology. And we are specifically designing 2-phase liquid cold plate technology, which is a plug-in replacement essentially to existing solutions today. So that we'll have a cooling technology that helps people deploying AI to transition from [rack mount] (ph) systems over time to 2 phase liquid immersion technology, but they'll be able to begin using our liquid cold plate solutions today to lower the cost of their cooling. And again, we don't use water in our systems. And so we are environmentally friendlier in those spaces. And we fully expect to see some really exciting opportunities there. We're currently deploying getting ready to deploy pilots with multiple compute OEMs on the AI side, testing solutions across a broad set of applications and configurations, such that we're able to really characterize the benefits for each of these OEMs. In the AI market, we believe the key thing is to partner with the compute OEMs, such that they use you as a reference what people should use for cooling. It dramatically decreases the sales friction and dramatically increases the potential demand for the product. And 2PIC is only one of the many solutions we intend to bring to the AI and digital infrastructure market. Over time, it's our intention to deploy a broader swath of solutions and services to such that they can leverage things like our power management capabilities, orchestration and other things. So we're very excited about what this market is going to bring. And we think 2PIC and our liquid cold-plate solutions are a great market entry vehicle for us as the AI technology industry continues to develop and as they advance into hotter and hotter processors.

Thanks, Fred. And I want to touch on a couple of things here. I know I have some moment of time. But one, I mean maybe just wrapping it all together. So you recently did the Auradine Series C funding round. I was wondering if you can maybe -- if you can disclose your ownership stake there and if you're interested in pursuing a majority position or potential vertical integration in the future, one? And then around this -- maybe just talk about how the tariffs imposed by the Trump administration are impacting procuring mining rigs. I think you spent about $136 million in CapEx in Q1 is split roughly 70-30 between rigs and infrastructure. So maybe just if you can combine all that into one. Thank you.

Sure. As it relates to Auradine, we own a little under 15% of the company today. And one of the reasons we were very involved in founding the company was because developing semiconductors is very expensive, and we felt it would be best done leveraging institutional investment dollars. And today, the investors in Auradine include people like Mayfield, Celesta, Samsung and many others. And one of the Board members on that serves together with me on Auradine’s Board is Lip-Bu, the CEO of Intel. So Auradine is very focused on solutions outside of Bitcoin mining as well. The Bitcoin Mining chip which were the predominant customer of is amazing, very power efficient and has a set of features that allow us to do all sorts of very special things in relates to -- in relation to dealing with the energy efficiency and how the chip operates towards a set target, whether that be hash rate, energy efficiency or maximum bitcoin production. Other things that Auradine are doing relate to, which they've announced related to AI infrastructure technology around network which is a super excited product. Auradine has already spun-off an AI security technology business. And if they keep doing the great job that they're doing, it's going to be a very valuable investment for MARA to have, but it's not something we'll make captive anytime soon. It just doesn't make sense for us. As you look at other aspects of tariffs and things like that, as we mentioned in the shareholder letter, we've now begun manufacturing our own miners. And obviously, we've been working to develop a supply chain around that. That enables us to optimize and have agility, if you would, and what we're doing is such that we're minimizing as much as we possibly can, the impact of tariffs. And if you think about a miner the vast majority of the cost of the Bitcoin Miner are the ASICs. And so again, the relationship we have with Auradine gives us the opportunity to keep our costs down, and control the supply chain such that we can ensure our own supply, but more importantly, we can minimize the impact of tariffs. As it relates to the infrastructure that we use, it is going to have an impact, obviously, steel, metal, things that you used to do containerized solutions are impacted by tariffs, and I think will be impacted in that area the same as anybody else would. But again, we have a very strong supply chain team that worked for us and especially with 2PIC and building our own technology solutions around containerization and infrastructure. We're going to optimize on a continual basis according to how tariffs go. I think my personal belief is that I think we'll see trade deals announced on kind of a continual basis here over the next months that will ameliorate a lot of the concerns for industries. But I think that there will continue to be a challenge with tariffs related to China. And I think that the more that you can source outside of China, the better things will be. And then just lastly on the topic of tariffs. As we continue to grow our business outside of the U.S., that's a very different tariff picture when you're buying miners or you're buying infrastructure and you're having that delivered somewhere outside the U.S.

Thanks Fred. So in March, MARA established a new $2 billion at-the-market equity facility. And I was hoping you could discuss sort of the strategic rationale for choosing an ATM program over taking on maybe additional leverage and outline your priorities for deploying these funds in '25 and beyond.

Stephen, when you think -- sure. When you think about the source of capital for this sector and for MARA in general, it's been predominantly again dependent over the last few years with some converts. And in our history, we have done some converts by in 2021. And put some leverage on the balance sheet and bought Bitcoin. And then we also did a 0% coupon last year to buy Bitcoin as you remember in the Q3, Q4 time frame. Our usage last year for the ATM was less than 50% of our total source of capital, as a result of that. So when we look at our capital profile and our sources of capital with the growth plans -- we look at what are the best source of cash and source of raising capital for that and responsibly look at the opportunities to raise the capital. We are not in the market every day. We opportunistically look at -- there are some very specific KPI-driven capital raise program that we have instituted over the last few years, and that has worked out well for us. Now when we think about raising capital, the capital has to be deployed in accretive projects. And that's why our threshold has been very high in terms of what -- how do we allocate our capital and what kind of rate of returns do we expect? And as we had talked about a few moments ago, the projects that we would expect to see in the future are going to squeeze more value out of the same investment that we had done in the past. The other thing that we have been focused on is another KPI, which is return on capital employed. And yes, quarter-over-quarter, it can fluctuate depending on the bitcoin price. But when you look at it on a long-term perspective, it is one of the most healthiest in the sector. And we take pride in that, but not only that, but we keep an eye on that as to what we invest our capital and utilize it we have prudent balance sheet management and preserve our flexibility to the source of capital available.

Thanks Salman. And last one for me. You have that full HODL approach that your Bitcoin treasury policy that you adopted in July of '24. You've seen other kind of miners in the market now starting to liquidate Bitcoin that's mine and so forth. Do you plan to sort of maintain the strategy through '25 and beyond and sort of what factors or market dynamics might prompt MARA, to reconsider or sell some of its monthly production?

Yes. I mean when you think about MARA, so the distinguishing feature that MARA versus look, investors have opportunities to invest in either bitcoin HODL, who is going out and buying Bitcoin from the open market at the spot price, which, by the way, is right now 103,000 or people like to invest in miners like us who are pure-play miners and they have -- for example, our cost per coin is one of the lowest in the sector. So any upside from the cost per coin to the Bitcoin price our stockholders benefit from that as the gold price kicks-in. What MARA is doing is a twin turbocharge strategy, if you like, right? If you are a car owner, you would know that if you drive a base model or if you drive a single triple engine, there are differences. And then when you drive a twin turbo engine, you see the benefit of both the sides, and that portfolio approach across the board is really attractive in our [humble view] (ph) from a long-term perspective. And we expect that to pay-off our stockholders from a long-term perspective as the Bitcoin price increases from here. So going back to the HODL approach. Last year, we went on full huddle strategy. We looked at the market conditions. We looked at the political ill winds and the macros, and it made a lot of sense to go out and HODL and we accumulate that Bitcoin that we haven't sold. And we also went out and bought Bitcoin by raising doing converts we represent coupon converted last year. And that was the best use of our capital at that point in time. Now we are not -- as you noticed that we are primarily a bitcoin miner. And -- but at the same time, our stockholders benefit from HODL, the second largest HODL in worldwide and as far as we know at over 48,000 coins today.

Thank you guys, very much. Appreciate it.

Thanks Steve.

Thanks Steve. Appreciate the questions. We’re going to pass on taking retail questions today as most of them are already addressed on today’s call. Thanks, everyone, again, for joining us. If you have questions that were not answered during today’s call, please feel free to contact our Investor Relations team at ir@mara.com. Thank you very much, and enjoy the rest of the day.

Thank you. This does conclude today's conference. We thank you for your participation. You may disconnect your lines at this time, and have a wonderful day.