Marathon Digital Holdings, Inc. (NASDAQ:MARA) Q2 2023 Earnings Call Transcript

Marathon Digital Holdings, Inc. (NASDAQ:MARA) Q2 2023 Earnings Call Transcript August 8, 2023

Marathon Digital Holdings, Inc. misses on earnings expectations. Reported EPS is $-0.13 EPS, expectations were $0.03.

Operator: Good day, ladies and gentlemen. Welcome to Marathon Digital Holdings Second Quarter Earnings Webcast and Conference Call. I would now like to turn the call over to your host, Chris Brendler, Vice President of Investor Relations. Please go ahead.

Chris Brendler: Thank you, Daryl. Good afternoon, and welcome to Marathon Digital Holdings second quarter 2023 earnings call. Thank you for joining us for our call today. With me on today’s call are Chairman and Chief Executive Officer, Fred Thiel; and our Chief Financial Officer, Salman Khan. Before we get started, I’d like to remind everyone that our prepared remarks may contain forward-looking statements, which are subject to risks and uncertainties, and that we may make additional forward-looking statements during the question-and-answer session. These forward-looking statements are subject to risks and uncertainties and actual results may differ materially. When used in this call, the words anticipate, could, enable, estimate, intend, expect, believe, potential, will, should, project and similar expressions as they relate to Marathon Digital Holdings are as such a forward-looking statement.

Please refer to our earnings release for a full recitation of our forward-looking statements. Investors are cautioned that all forward-looking statements involve risks and uncertainties, which may cause actual results to differ materially from those anticipated by Marathon at this time. In addition, other risks are more fully described in Marathon’s public filings with the U.S. Securities and Exchange Commission, which can be viewed at www.sec.gov and ir.mara.com. Finally, please note that on today’s call, we will refer to certain non-GAAP financial measures in which Marathon excludes certain items from its GAAP financial results. Please refer to our company’s periodic reports on Form 10-K and 10-Q, and to our website for a full reconciliation of these non-GAAP performance measures to the most comparable GAAP financial measures.

We will begin today’s call with prepared remarks from Fred and Salman. After their comments, we’ll be going through some of the more popular questions from investors before transferring to a live Q&A with our covering analysts. And with that out of the way, I’m going to turn the call over to Fred to kick things off. Fred?

Fred Thiel: Thank you, Chris. After a solid start to the year, we continued to make progress towards our primary 2023 goals, which include energizing our fleet to reach our 23 exahahash target, while at the same time, increasing our operational efficiency. As our recent results demonstrate, we’re near to completing our first goal and we’re making consistent progress on the second. During Q2, we increased our operational hash rate 54% from 11.5 exahashes on April 1, to 17.7 exahashes on June 30. Those of you who saw the July production report we published last week will have noticed that our operational hash rate is now approximately 19 exahashes. In 2023 alone, the team at Marathon has more than doubled our operational hash rate and positioned us as the largest publicly traded self-miner on an installed hash rate basis.

We’ve grown hash rate this year by successfully working with our various hosting partners to energize new facilities. During the quarter, the facility at Ellendale, North Dakota was energized and officially became our largest operational Bitcoin mining site. While growing hash rate has been our primary goal, it has been important for us at Marathon that we do so in a sustainable manner. Since I became CEO, we have made a concerted effort to transition Marathon to more sustainable sources of power, and I’m pleased to report that we have been successful in that endeavor. In less than two years, we’ve evolved from a company that relied on a coal plant for its energy to one that is mostly powered by renewable energy. During the first six months of this year, approximately 58% of the energy our miners consumed was generated by renewable and sustainable sources.

While we expect this mix to fluctuate depending on wind conditions, curtailment and other factors, we remain committed to 100% carbon neutrality in 2023, with the balance offset by renewable energy credits or similar instruments. Operationally, we hit another milestone this week. As of today, we have officially installed 23 exahashes of miners, and energization of the final tranche to reach 23 exahahash of operating capacity appears to be imminent. I’m pleased to report the final tie-in to the substation in Garden City is complete; therefore, barring any unforeseen delays, energizing this site and reaching 23 exahashes of operational capacity should be imminent. Because we’ve been growing our hash rate faster than the rest of the network and simultaneously improving our uptime, we have continued to increase our Bitcoin production.

We produced a record 2,926 Bitcoin during the second quarter, which is a 33% increase from Q1 of this year. And last month, we won a record 179 blocks, which is a record for Marathon and indicates that our improvements have continued into the third quarter. In a few minutes, I’ll provide an update on our expanding capabilities and evolving growth strategy, but first I’m going to turn the call over to Salman to discuss our financial results for the second quarter. As a reminder, Salman joined us as CFO in June, and he has quickly become a valuable member of our team. He brings a wealth of valuable experience from several industries that are directly pertinent to Marathon, including technology, renewable energy, oil and gas, and Big Four public accounting and consulting.

With that, please join me in welcoming Salman to his first earnings call with Marathon. Salman?

Salman Khan: Thank you, Fred, and thanks for the warm welcome. I look forward to meeting and engaging with our analysts and investors over the coming months. For now, let’s focus on the results for the quarter. Improving Bitcoin prices, increased production and accounting gains on Bitcoin sales were the main drivers of our operating results for the second quarter. The company recorded a net loss attributable to common stockholders of $21.3 million, or $0.13 per common share in the quarter, compared with a net loss of $212.6 million or $1.94 per share in prior-year quarter. This represents a $191.4 million favorable variance in net loss year-over-year. Total revenues for the quarter were $81.8 million, well above our prior-year revenues of $24.9 million, representing a 228% increase year-over-year.

The company’s Bitcoin production at 32 Bitcoin per day was 314% higher, which was partially offset by a 14% lower average Bitcoin price compared to last year’s quarter. Our hosting and energy costs for the three months ended June 30, 2023 were $55.2 million compared to $16.7 million last year. The increase was due to increased Bitcoin production, partially offset by lower production cost per coin and one-time exit costs for our Hardin, Montana facility last year. Total cost of revenues, which includes depreciation, amortization was $92.5 million compared to $41.4 million last year increasing by $51.1 million, or 123%. Depreciation and amortization this quarter was $37.3 million increasing by $12.6 million compared to the same quarter last year.

We also had one-time accelerated depreciation of our Hardin, Montana facility for $9.4 million last year. As Fred mentioned, we have grown our mining rigs significantly since last year. And as a result, increased our energized hash rate from 0.7 exahashes in the second quarter of 2022 to 17.7 exahashes in the same quarter of 2023. Gains on sale of digital assets was $23.4 million in the current year quarter. And in the prior year, we had a loss of $14 million from losses on digital assets loan receivable. Impairment of digital assets improved by $123.2 million year-over-year as Bitcoin prices fell significantly in the prior year’s quarter. Lastly, you remember last year we had a Bitcoin investment fund with a $79.7 million loss and a $54.1 million gain on sale of equipment, both of which did not exist this quarter.

I want to note that the company’s non-GAAP total margin, excluding depreciation and amortization, was $26.5 million this quarter, representing an $18.3 million, or 222% increase when compared to the same quarter last year. This is representative of our increased hash rate and our strategy to systematically improve efficiency. Adjusted EBITDA improved to $25.6 million versus $167.1 million loss in the prior-year period. The drivers of adjusted EBITDA include the previously mentioned accounting impacts and total margin improvement excluding depreciation and amortization, partially offset by a $10 million year-over-year increase in general and administrative expense to support the growth of our business. General and administrative expenses, excluding stock-based compensation, we’re $16 million in the current period as compared with $4.3 million in the prior-year period.

This increase in expenses was primarily due to higher cash compensation and benefits, professional fees, property taxes, charitable contributions and insurance costs. I’m pleased to report we now have over 40 employees, up from just 17 a year ago, as we have built out our finance, strategy, communications, legal and human resources teams and opportunistically added talent across the organization. Turning to our Bitcoin Holdings and cash position. Cash and cash equivalents totaled $128 million at June 30, 2023, an increase of $15.5 million compared to last year. Also, at June 30, 2023, we held approximately 12,538 Bitcoin with a carrying value of $234.4 million on the balance sheet, with the fair value of our holdings approximately $382 million at quarter end.

The company’s combined balance sheet of unrestricted cash and cash equivalents, and fair value of Bitcoin was approximately $495.7 million. We sold 1,854 Bitcoin during the quarter, realizing cash proceeds of $51.3 million. These proceeds were utilized to fund operating expenses during the quarter, including cost of revenues for energy hosting and other cash operating expenses. During the quarter, we generated $65.5 million from at-the-market sale of equity, which we intend to use for growth capital. The combined cash and cash equivalents Bitcoin and access to our – at the market facility provides us ample amount of liquidity and optionality to execute our strategy. So all in all, a good quarter for the balance sheet as we maintain our strong cash position while increasing our Bitcoin Holdings.

As mentioned in our July production report, our unrestricted cash balance at July 31 was $115.1 million, and we held approximately 12,964 Bitcoin with a fair value of $379 million. We expect our future Bitcoin Holdings will generally increase, but will fluctuate depending upon operating and market conditions. We intend to add to our Bitcoin Holdings primarily, through our production activities. And we will also continue to sell Bitcoin, as a means of generating cash to fund monthly operating costs, and for general corporate purposes. And that completes my update. So I’ll turn it back over to Fred who will take a little bit more about – who will talk a little bit more about our operations and our ongoing plan. Fred?

Fred Thiel: Thanks, Salman. While all Bitcoin miners fight to win blockchain in Bitcoin. Marathon has always done things differently. We’ve achieved a number of first for the industry. We’ve set the pace for growing hash rate to improve our Bitcoin production for navigating capital markets to efficiently scale the business, for operating behind the meter of wind farms to improve our mix as sustainable energy, and for geographically diversifying our operations domestically, and internationally to reduce concentration risk. Now we’re continuing this trend by vertically integrating our technology stack, which can potentially reduce latency in our operations. If you can reduce latency then in theory, you can operate more efficiently, and therefore potentially win more blocks.

Efficiency is particularly important as you move closer towards next year’s having from the block subsidy will be reduced by 50%. Because Bitcoins difficulty rate dynamic, someone will always make mining I’m sorry, make money mining Bitcoin. But what matters, is that you are more efficient than the average miner. And if you’re on the low end of the cost curve that you can keep operating when others cannot. And we believe the technology is one of the keys to improving efficiency and operating even in those toughest of times. Marathon is the only Bitcoin miner that directly controls or influences each aspect of the mining tech stack. Starts at the top of the mining pool, which is the orchestration layer, between the miners and the blockchain, and it runs all the way down to the ASIC.

Over the past few months, we’ve made several upgrades to MaraPool, which seem to have contributed to an improvement in the number of blocks won, and therefore, Bitcoin earned. However, developing our own proprietary firmware for our miners is making an even larger impact. There are several companies that sell firmware. And given our experience in our technological expertise, we thought we might be able to build a superior solution. And so far the feedback indicates that we have. We recently benchmarked our proprietary third we’re against some of the leading alternatives. And so several advantages in performance, protections and security. So far our firmware seems to have demonstrated better curtailment times that are overclocking, improved temperature and voltage protection and encrypted communications with the mining pool.

And it allows for security audit, which is a very valuable tool for us as a publicly traded company. Moving down this tech stack to the hardware. We’ve been heavily focused on upgrading our fleet of S19 XPs to not just grow hash rate, but to also reduce our energy expenditure per hash our mining fleet now operates at 25.3 joules per terahash which makes it one of the most efficient fleets in the industry. At the infrastructure level we’re clearly demonstrating our competency in designing and building immersion cooled mining farms. In February, we signed an agreement to jointly build a 250 megawatt facility in Abu Dhabi. For reference, the climate in Abu Dhabi is one that is extremely hostile towards Bitcoin miners. The temperature often is above the typical shut-off temperature for Bitcoin miners in air meaning over 40 degree C very high humidity and it’s a very demanding environment.

The sites in Abu Dhabi where Marathon today is partner and its joint venture were designed and built and are operated by team trained by Marathon. And it really exemplifies the ability that Marathon has in building these sites. And as a matter of fact, our pilot site operated for over 100 days without human intervention, which shows how sophisticated our operations can be. So, we clearly believe that this demonstrates our ability to now, become truly a leader in the owned, and operated space as well as third-party hosted. So the whole 250 megawatts in Abu Dhabi is expected to be online before the end of this year, but already in July, first of the two sites began hashing and produced for 17 Bitcoin of which we received 3.4 Bitcoin due to our 20% ownership in the project.

Even the hash rate that’s currently online has already proven to be helpful to the region as a load balancer and preliminary reports have shown the site’s ability to dynamically ramp up and down during the day when grid energy demands change. This type of flexible demand response has an excellent case study for how Bitcoin mining can benefit utilities. It also proves, that we can now mine Bitcoin successfully in parts of the world that were inaccessible even a year ago, given their climates. While we’ve not announced any significant growth plans beyond 23 exahashes. Our intention is to continue to grow and maintain our leading position in this space. Our scale, operational expertise, flexible approach to implementing the right business model for a given deployment, and our ever increasing portfolio of technological innovations are all key to drive rapid expansion of efficient operations.

With the having coming up next year optimism may not be the dominant emotion for every Bitcoin miner. And many in our industry today are focused on how they’re going to survive, which is not easy to do, given how dynamic this spaces. But at Marathon, our focus is not just staying alive, but thriving. Even in the most challenging times. We remain focused on conscientiously growing our hash rate, and consistently improving our technology and remaining a leader in this industry. We’ve been setting the pace in this industry for the past three years, and we have every intention to keep that trend going. This industry is still young, it’s full of potential for those who are willing to shape it and we therefore believe the most exciting years for Marathon and for Bitcoin lie ahead.

Chris?

Q&A Session

A -Chris Brendler: Thanks, Fred. At this time, we’re going to commence the Q&A section of today’s call. I’ll start by answering some of the most popular questions. Submitted by investors through our Q&A platform. Our first question comes from CK who asks, regarding Auradine and MARA; number one, what percent of Auradine does MARA own? Two, how should we think about the pricing/benefit that having a stake in Auradine will provide to MARA as you continue scaling hash rate? And three, is Auradine providing AI capabilities to MARA within the tech stack? Great question, CK. I’m not sure how much of that was disclosed at this point. Fred, can you add a little color on Auradine?

Fred Thiel: Sure, great question. So Marathon today owns about 12% of Auradine, so we’re a minority shareholder. There are large venture capital firms that own a much larger percentage, and then the founders of Auradine own the balance. I think you should think about the reasons why we made the investment rather than talking about price advantages and things like that. What we really were looking for was to find the ability to have a miner built that could really operate in all of the challenging environments we needed be able to ramp up and overclock, be able to ramp down and underclock, be able to operate in immersion, single phase, dual phase immersion, have very efficient operations, but more importantly, be able to be intelligent.

And one of the key things with the Auradine miners that the team in Auradine has clearly been able to achieve even beyond the expectations that Marathon initially had is to create a miner with a total cost of ownership that is significantly lower than most of the industry. And that is so important as we move into an environment with very dynamic energy pricing and also the halving. And dynamic energy pricing is critical because these miners essentially can operate in a way so you can set a price target of energy or a price target of Bitcoin, if you would – hash price, if you would, for Bitcoin and a variety of other parameters. And the miners will automatically adjust to operate within those parameters, scaling up their capacity, scaling down their capacity.

It is a game-changer for the industry, which again means that these miners can operate in large fleets managed from the cloud platform that makes it very unique. So that’s what we’re most excited about as regards to Auradine, and we’re very excited to be able to start incorporating those systems into our fleet.

Chris Brendler: Great. Our next question is from Tariq who asks, what’s your future projections for mining production? Are there any things you’re focusing on besides Bitcoin mining to also bring more revenue to your company? You want to take that one as well, Fred?

Fred Thiel: Sure. So while we haven’t disclosed kind of what we’re doing outside of pure Bitcoin mining, and I would just say stay tuned to that, we’re going to continue to grow our operations. As I said in my prepared remarks, we intend to continue to remain a leader, and we’ll continue to grow capacity at a conscientious rate over the coming years. We believe that it’s our duty to be one of the leaders in this industry and supporting and securing the Bitcoin blockchain and supporting the industry and the infrastructure around it. So as you look towards the future, expect to see Marathon developing technologies that will help build the infrastructure for this industry to really thrive long term.

Chris Brendler: That’s great. Next we have a question from Andreas who asks, the halving is rapidly approaching and with it, a looming bull run. What are your plans upon this rapid rise in Bitcoin price? What are you doing to separate yourself from other mining companies? So I appreciate your optimism, Andreas, but I don’t think we’re going to comment on Bitcoin price predictions at this point. Fred, maybe you can provide some more color on how we are approaching the having and how we are differentiating ourselves.

Fred Thiel: So having, is obviously a recurring event in this industry and all miners are very focused on making sure that there’ll be able to operate profitably, not just at the having but beyond and through the next cycle – until the next having in 2028. So, as Chris mentioned, we’re not going to comment on where we think the price of Bitcoin is going to be. But we’re obviously very focused on optimizing our operations that our cost to mine Bitcoin, will allow us to weather any sort of storm here as we get towards the halving where Bitcoin price may fluctuate up and down kind of greatly. As Bitcoin price moves up, obviously there will be an increase in hash rate because more miners will want to be putting on more hash rate.

If the price of Bitcoin doesn’t grow into the halving, then you may even expect, as some analysts have predicted, the hash rate to actually decrease post-halving. So we’ll just all have to look and see how that is, but we’re very focused on just being the most optimized miner from a cost perspective. And while we still have some ways to go, we’re very focused on it. And as our mantra has been this year, it’s energize and optimize, and so we’re very focused on doing both at the same time.

Chris Brendler: That’s right. The next question comes from Bert F. who asks, why should an individual invest in MARA as opposed to Riot or buying Bitcoin itself? What do you see as the benefits of MARA compared to other investing options? Great question, Bert. Allowing to Fred as well.

Fred Thiel: So I think Riot runs a great operation. I think the challenge with Riot is they’re not a Bitcoin pure-play in the sense that they operate a third-party hosting business where they rent out capacity to third parties, and then they also are a self-miner. They are also highly concentrated to a particular area in Texas, which makes them very susceptible to either regulatory changes in Texas, ERCOT changes in Texas, climate, et cetera. So I think you have to look at each miner’s operating model. We operate a very diversified portfolio. We operate some third-party hosted sites. We operate some owned and operated sites. We operate in Texas. We operate in North Dakota. We’re now operating in UAE. And we’ll continue to expand internationally and continue to diversify our portfolio of sites and move to a greater balance of owned and operated than we have today.

So we offer kind of more of a diversified play in that regard. We are also a Bitcoin pure-play in that we not only mine Bitcoin, but we hold a lot of Bitcoin. And that generally tends to show itself in how you see our stock react to increases in the price of Bitcoin. So I think Riot, Marathon, two very different companies and operating under different models. I think that, yes, if you compare Marathon to just buying spot Bitcoin, I would recommend you look back at Bitcoin price movements and Marathon stock movements and see and you’ll likely find that there were opportunities this year, for example, and I think even on a year-to-date basis where Marathon has outperformed Bitcoin. So that may be a reason to invest in the stock.

Chris Brendler: Okay. And then we have one last question from our – or time for one last question from our platform this afternoon, and we’ll actually combine two of the most popular ones into one and try to hit them both. James R. asks, are you thinking about or planning a reverse split? And then Xavier T. asks, why do you continue to wish and try to dilute the shares of your shareholders? Why not do a split, take a loan or do something similar to advance shareholder value instead of always trying to create more shares and dilute the value? So perhaps we can provide some color on our strategy for you as in equity and growing the company. Salman, maybe you can start and, Fred, you can finish it up.

Salman Khan: Sure. Happy to do that. Thanks, James, and Xavier for asking the question. The way we see as – from an equity and a capital stack standpoint is our eyes are focused on shareholder value creation, so any decision that we make, we will always think about shareholder value creation first. And when it comes to whether it’s debt or shares, we already have convertible debt on our balance sheet, which is low interest rate. And certainly, that there is evidence that we have the capital stack in the balance sheet that creates the value for our stockholders. In terms of growing and raising capital and the strategy to grow the exahash, when we look at the growth here, we’re not growing just for the sake of growing. We’re growing with the intention of creating value for the stockholders.

As long as it’s accretive to our existing stockholders, we will continue to raise capital to grow from that perspective. So hopefully, that answers your question. In terms of a reverse stock split, I’m not certain if I understood the question. We don’t have the intention at this stage, given our stock price and given our company size, we don’t see a value in a reverse stock split at this stage. Hopefully, that answers your question. Fred?

Fred Thiel: Yes, I would just add the comment that if you looked historically, splitting a stock is something you do when your stock price is very, very high. I mean, Apple did it when their stock was between $600 and $700 a share because it makes it look like a share is more affordable. And they dropped their share price down to – they did a seven split, so they dropped it down to about $90. Trading in the $15 to $20 range where we trade, it doesn’t make sense to do a split. And that doesn’t create any shareholder value. And doing a reverse split, you only really ever do if your stock price gets below $1. So I don’t see any value in doing that. But just as you look at our growth, every exahash that we add, adds shareholder value.

Because it produces a certain amount of Bitcoin and that amount of Bitcoin, a certain percentage of that we’ll continue to hold on the balance sheet. And that continues to add shareholder value. We have been selling Bitcoin to cover our operating expenses this year. So we’re not diluting our shareholders to maintain operating expenses in that degree. So I think our growth plans are going to be very consistent with doing things that drive shareholder value accretion.

Chris Brendler: Okay. On the interest of time, we’ll have to wrap up this section of the Q&A. Again, we really appreciate all the questions and interest, and thanks for submitting those questions. I’m now going to turn the call back to the operator to open the line for questions from our covering analysts. Daryl, back to you.

Operator: Thank you. We’re now going to open the call to questions from Marathon’s covering analysts. [Operator Instructions] Our first questions come from the line of Lucas Pipes with B. Riley. Please proceed with your questions.

Lucas Pipes: Thank you very much, operator. Good afternoon, everyone. Thank you for taking my questions. My first one is also on the halving side. So with it rapidly approaching, I’m trying to get a sense for the flexibility you have. Just for argument’s sake, if a given data center is not economical on a variable basis, do you have the flexibility to curtail miners there temporarily at least, or were there kind of take-or-pay commitments if you were to take that step that you would still have to pay for a given amount of power? Would really appreciate your color on that dynamic. Thank you.

Fred Thiel: Thanks, Lucas. It varies by site, and I’m not going to go into each site one by one. But it depends on if at the site we control uses the PPA or we’re doing pass-through power. And that happens to be the case in some of the sites there are other sites that have very low cost fixed power and other sites that just have fixed pricing, which is power and hosting. So it’s a broad mix between the three.

Lucas Pipes: That’s helpful. Thank you, Fred. And just to follow up on that, kind of as a percentage of exahash rate today, and assuming installed exahash rate, roughly what percentage would fall under fixed commitments and which ones would be more flexible?

Fred Thiel: I think you would find of the installed, about a third today is flexible. And the balance is most probably – I’m not going to guess, but it’s most probably similar, 40% fixed and then the balance is very low cost fixed.

Lucas Pipes: Really appreciate that, thank you very much for that. Two quick follow-up questions, not directly related, but first in terms of the utilization rate on the operating hash rate. What are kind of your expectations on that metric going forward? We calculated in-house, but would appreciate how you approach that question. Then, yes, somewhat not directly related, but Garden City, you mentioned you expect to start up imminently because imminently is pretty clear, but if you could maybe share your update on the site and when you would expect it to turn online, I would really appreciate it. Thank you so much for your color.

Fred Thiel: Yes, so the physical tie-in at the site is complete. And the last documents between our hosting provider and the energy provider are being passed back and forth with the final red lines. And as soon as that document is signed, they start energizing. So that’s the last step. There are no other approvals or any other regulatory issues holding that site up from going on. So our hosting providers publicly said they expect it to happen imminently. We believe the same based on the current status of their discussions with the power company. And that will ramp in kind of 20-megawatt steps over a short period of time.

Lucas Pipes: Got it, so this would kind of ramp up over the course of 10 weeks or so?

Fred Thiel: Most probably somewhere between six weeks to 10 weeks.

Lucas Pipes: Great. And then in terms of utilization rates more broadly?

Fred Thiel: You mean uptime? So help me understand what you mean by utilization.

Lucas Pipes: Yes. Sorry, with that, I mean, yes, uptime. Obviously like that Garden City miners don’t run, that’s understood. But kind of in terms of the operating base, kind of what uptime is a good benchmark.

Fred Thiel: So it’s seasonal, as you’re well aware. Texas has a lot of curtailment in the summer. And so operating in Texas, uptime is going to be lower in Texas. In North Dakota, by contrast, we’ve been operating well over 93%, 94% even in the summer. And so that site has been operating with very high uptime. So I think you have to kind of look at the West Texas sites as being very seasonal. You’re going to be in the mid ’90s in the cooler times of the year. And depending on curtailment and temperature issues, you may be in the 60% to 75% during some of the hotter months in the summer.

Lucas Pipes: Super. Hi, Fred and team, I really appreciate it and best of luck.

Fred Thiel: Yes.

Operator: Thank you. Our next questions come from the line of Chase White with Compass Point. Please proceed with your questions.

Chase White: Thanks. Good evening, guys. So how should we think about aggregate power costs across the entire portfolio as you energize Garden City and that site starts to come online? And then I have a follow-up.

Fred Thiel: Yes. So what complicates this is that the APLD sites, we pay essentially a fixed price for power and hosting. So power isn’t broken out net cost, it’s power and hosting. So that makes it a little bit harder to give you just the power costs. But I think you could look, generally speaking, across sites and you would find us – if you include power and hosting across all the sites, you’re going to be in the kind of $0.06 range per kilowatt.

Chase White: Got you. And switching gears a little bit. I mean, are you able to provide us any additional details about the economics for the Abu Dhabi JV?

Fred Thiel: It’s an 80-20 joint venture. We’ve just started production, so we don’t even have first month’s financials yet, so I couldn’t tell you. But what I can tell you is it’s a fixed price – sub-$0.04 fixed price energy, and then we split the operating costs 80-20.

Chase White: Got you, so sub-$0.04. And is that – and that’s fixed just all the time. Perfect. And then any updates on the recoveries from the Compute North bankruptcy?

Fred Thiel: None to mention, no.

Chase White: Got it, thanks.

Operator: Thank you. Our next questions come from the line of Jon Petersen with Jefferies. Please proceed with your questions.

Jon Petersen: Great, thank you very much. I’m just curious on the exahash here – or the XPs that you’re putting into place right now at today’s economics. What do you estimate like the payback period is on what you paid for those machines at today’s economics?

Fred Thiel: Well, those machines were significantly discounted by the time we finally paid for them, right? So when we ordered them, it was at a high price. And by the time we ended up paying for them, it was significantly below that as can be seen by the limited amount of CapEx we’ve had to do to pay off those, so I think you’re going to find it’s somewhere between 13 months and 18 months, most probably.

Jon Petersen: Okay. All right, that’s helpful. And then I know this is a hard question to answer, but I was curious in terms of the halving next year, do you have any estimate of what percent of the exahash that’s online right now would come off post-halving?

Fred Thiel: Well, you tell me what the price of Bitcoin is going to be, and then you can touch it yourself.

Jon Petersen: Let’s just pretend it’s today, if it’s today’s price of Bitcoin and today’s exahash. I know there’s a lot of levers to figure this stuff out.

Fred Thiel: I mean, if we’re talking about a price of $30,000 for Bitcoin, that would mean essentially at the halving, you’re talking about the equivalent of $15,000, which would be right around our marginal cost to produce. So don’t know we would shut off necessarily.

Jon Petersen: I didn’t mean any – sorry, I didn’t mean any of yours. I meant the market overall, like do you have any sense of how much of the total network hash rate might come offline.

Fred Thiel: Well, look at it this way. Our fleet operates at roughly 25 joules per terahash. The average for the industry is somewhere in the mid ’30s. Your colleagues are estimating 20% to 30% would come off.

Jon Petersen: Yes, okay. All right, that’s all –

Fred Thiel: So if you’re in the lower third of the miners, from an efficiency perspective, I think you’re going to do fine. I think it’s, if you’re above – if you’re average or above, you’re going to be challenged.

Jon Petersen: Yes, that’s fair. All right, that’s all for me. Thank you.

Operator: Thank you. [Operator Instructions] Our next questions come from the line of Kevin Dede with H.C. Wainwright. Please proceed with your questions.

Kevin Dede: Hi, Fred. Thanks for taking it. I was curious about your software integration. How much, progress have you made on that, how much have you implemented, what sort of efficiency gains have you seen, can you talk to that at all?

Fred Thiel: I can talk a little bit to it. So we’re running. We’ve just transitioned to our third generation of pool software, which is fully kind of developed Marathon and that is the one that’s been driving a lot of efficiency gains. I want to say how much, but let’s just say that it’s been driving efficiency gains. Our firmware is running in a subset of miners and we continue to expand that across the fleet and our expectation is to have the vast majority of the fleet transition to our firmware here over the balance of this year. That far more gives us additional efficiencies, which I think will be very beneficial. The immersion technology that we’re running in UAE and we’re also running a tiny bit of it in some other places in the U.S. that we’ve done for testing purposes was developed by a third party with some direction for Marathon and we’re very focused on expanding the development efforts, we’re doing in immersion to get even better efficiencies and the efficiency in the immersion, have more to do with being able to operate not just an extreme environments but also require fewer touches by human beings because that’s where you get the operating efficiencies in immersion is if you don’t have to touch it and you can have fewer engineers on-site.

And so we’ve been very focused on putting a lot of instrumentation into the system. So that they can be managed remotely. And if you recall my kind of comments regarding Auradine to a prior question, with the capabilities that those miners bring over time we’ll have the ability to really manage a fleet from a profitability standpoint where the machines and the orchestration layer really manage themselves if you based on profit targets and cost points. So that’s kind of what we’re very focused on. And that’s why we felt it was so important to make the investment in Auradine so we could have the ability to influence the design of the miner all the way down to the ASIC level. So we can get specifically what we felt was the right solution to the marketplace for our needs.

Kevin Dede: So Fred to load the firmware and Auradine software, do you need to pull the miner from the RAC take it to the shop.

Fred Thiel: Well, so the Auradine software will run on the Auradine miner. It won’t run on the Bitmain so let’s be clear about that.

Kevin Dede: Okay.

Fred Thiel: Below the software we do over the air software upgrades.

Kevin Dede: Okay. And how flexible do you think the software development is and will be to changes in your fleet.

Fred Thiel: So far it’s been very flexible. Yes, got think about it this way. Our fleet consists of two types of machines today S19 J PROs and S19 XPs predominantly. As we add other machines will add different versions of firmware as it makes sense. So if a machine comes with the firmware that’s as good as ours and has the same capabilities, there is no reason for us to change the firmware and the machine. It just so happens that Bitmain’s machine don’t come with particularly good firmware when it comes to the ability to overclock underclock set 10 points, things like that, it’s just, it’s very minimal. And the other third-party firmwares might be good at one or two things that are not good at everything. And so that was why we made the decision to develop.

Kevin Dede: Okay. Thanks for the detail. I appreciate you taking my questions. Thanks, Fred.

Operator: Thank you. Our next questions come from the line of Lucas Pipes with B. Riley. Please proceed with your questions.

Lucas Pipes: Thank you, operator. Thank you for taking my follow-up questions. First, Fred can you remind us about the fleet at Garden City. Are those most of —

Fred Thiel: All XPs.

Lucas Pipes: All XPs? Got it. So kind of we should expect the efficiency to tick up as Garden City turns online.

Fred Thiel: Yes. So I think we’ve said previously that our once fully deployed, we would be at about 22 or 23 joules per terahash and we’re at 25 today. So that gives you kind of feel for that.

Lucas Pipes: Got it. Very helpful, thank you for that. I know you had said it before. I just wanted to make sure I remember it right. And then Fred, on M&A, do you have thoughts on what could make sense here, the industry obviously kind of came through some of the most challenging times. And I could imagine there are players might be looking for liquidity. And so as a leader, you might have a role to play, I would appreciate your thoughts on that.

Fred Thiel: Sure. I think I’ve said a number of times that M&A in this industry is challenging because the replacement cost tends to be about the same as the acquisition cost when you think about most miners because the hardware has a fairly short life and depending on the age of whatever fleet the miner is running, you always have to weigh – what am I going to pay for these machines versus buying new and as everyone’s aware today, the cost of new miners is a rock bottom level you’re paying single digit dollars per terahash for S19 J PRO class machines and XPs are in the mid-’20s somewhere. So it’s almost better today to, if you’re going to acquire somebody, it’s got to be somebody that has an absolute amazing PPA or access to very low energy costs and a facility where you can essentially pull out old machines and put in brand new super-efficient machines.

There’s got to be a reason why the company is for sale typically right. So either the miners in trouble or they will be in trouble with the having and therefore the trying to sell themselves now. And I think there’s certainly a lot of activity in the marketplace right now with people who are trying to solve that issue. Because these miners that are in trouble today are going to be an even more trouble come having. And so I think they’re going to be even more opportunities come next year. But if I were to advertise for a minute. What would Marathon want to buy in the way of miner, it’s somebody with great access to very low cost power and great infrastructure where they’ve yet to put in any miners. And so we can build the site and we can equip it.

Lucas Pipes: Thank you, Fred for the perspectives. Maybe one last follow-up question. It’s a little bit of a higher level question, but I would appreciate your thoughts. Fred kind of when I look back over the history of Bitcoin mining every couple of years a new mining that came out, fairly relentless march upwards in the efficiency. I wondered what are you seeing today as the pace of technological advancement stable, decelerating, accelerating and with some of the developments you’re seeing on the AI front with very efficient ships is that a positive or possibly a negative because it increases the overall demand for new computing system. So kind of double at – two-pronged question here. I would appreciate. Thank you.

Fred Thiel: Yes, so in the Bitcoin mining space as a apart from AI space – in the Bitcoin mining space we went through a pretty significant increase in efficiency with the advent of the XP taking us from basically what was an industry average of around 30 joules per terahash down to about 21 joules per terahash. So that’s a pretty significant increase in efficiency – 30-ish percent. What we’re seeing now is a raft of new machines coming from whether it’s micro BT, whether Auradine miner, it was a company that were sharing the machine at mining disrupt which are all in that 20-ish joules per terahash. But if you overclock them or underclock them et cetera, you can get down to there – if you underclock you can maybe down into the high-teens.

Getting from there to the next level of efficiency in Bitcoin mining is going to require a pretty big step. I fully expect most of the – especially people like Bitmain working on next generation technologies and so it wouldn’t surprise me if we heard something new from them before the end of this year. But I think it’s going to be the returns on efficiency are going to start becoming smaller from a miner hash rate perspective where the real efficiency gains are going to come is in the software, not just in the miner meaning the ability to totally tune the operation of the miner to either a target hash price or energy cost or global hash rate et cetera. So the miner can literally tune itself to operate. Do I overclock? Do I underclock? What do I do so that I can maximize the amount of Bitcoin Mining and then you start getting into a lot of AI-driven cloud orchestration for fleets of miner.

So large operators like ourselves could reap very large benefits from that more so than increasing the efficiency of the hardware level by one or two joules per terahash. So that’s higher level.

Lucas Pipes: Yes, sorry, go ahead.

Fred Thiel: I was just going to say, as you look at the AI side of this business meaning Bitcoin miners moving to running AI compute versus mining Bitcoin, I see there are a number of miners who are actively focused on that Hut 8, Applied et cetera are very focused on that and what that does as I look at it relative to Bitcoin Mining is it means those companies will have less capital to invest in growing hash rate which means global hash rate from players that go that route won’t be growing. I think we’re going to continue to see the growth in global hash rate is from sovereigns who want to mine Bitcoin and I think you’re going to continue to see the trend of sovereigns mining Bitcoin as a way to potentially use Bitcoin as a reserve asset. I think it’s going to be a much more important trend to this industry going forward.

Lucas Pipes: Thank you. Thank you, Fred. A couple of interesting things there. On the comment regarding a smaller advance in the machine efficiency, where would you put the next generation at – 15 joules, 10 joules what would be your best guess today?

Fred Thiel: So the next big jump will be a geometry change. So for example Bitmain’s XP is a 5-nanometer, if they were to jump to 4, you may see it come down to 18 joules of terahash, if they jump all the way down to 3 which should be expensive, you may see it come to 17 maybe 16, but again that’s going to be a very expensive miner because of the cost of 3-nanometer wafers today. So that’s why I keep thinking the real benefits are going to come in software, as opposed to hardware over the next two to three years, and then we’ll have to see advancements in compute and in technologies typically take a while to flow through. And at the end of the day, Bitcoin mining is a fairly small volume industry for technology companies. When you think about it, they’re not hundreds of millions of miner sold a year.

And so it’s hard to justify a development effort. And I think it’s the type of market where it’s typically student for one or two players to kind of dominate the hardware side of it. And then the small marginal players just after you got little things they can.

Lucas Pipes: Very helpful. Lots of food for thought there. Fred, I really appreciate your perspective.

Fred Thiel: Yes. Thanks.

Operator: Thank you. At this point there are no further questions. I’m going to turn the call back over to Chris Brendler for closing remarks.

Chris Brendler: Thank you, Daryl. And just a quick clarification on the question around our fleet efficiency when fully deployed, we actually put in our press release on July production last week that we’re expecting 24.2 joules per terahash when fully deployed. So with that, thank you all for your time today. Do you have any questions that were not answered during today’s call, please feel free to contact our Investor Relations team at ir.mara.com. And with that thanks and enjoy the rest of your day.

Operator: Thank you. This does conclude today’s teleconference. We appreciate your participation. You may disconnect your lines at this time. Enjoy the rest of your day.