Hey miners, welcome or welcome back to the channel MCN Money, the official home of power mining analysis. In today's episode, Anthony Power and I are pleased to welcome Daniel Roberts back to the program. He's the co-founder and co-CEO of IREN. They just came out with their Q2 earnings, and we couldn't be more excited to go through the results in today's episode. Before we get into it, take a second, hit the like button. You guys, big help to myself and the channel, specifically in getting this content to other people like you who may find value. If you're not already subscribed to MCN Money, feel free to join and let us know in the comments section below if you're currently holding shares of IREN, what you thought about their quarterly results, and your outlook for the remainder of 2025. Now, with that being said, let's get into today's video.
Okay guys, so that's right. Today's video, we've got Daniel Roberts, the co-founder, co-CEO of IREN, back on the program. It's been a few quarters since we spoke with Dan directly. They just came out with Q2 earnings last night. Anthony and I just finished up listening to the earnings call and could not be more excited for the discussion. So Dan, great to see you. Thanks for making time.
No, thanks for having me back.
You bet. Hey, one of the first questions I wanted to ask: this line you gave us in 2024 stuck with Anthony and I throughout the year. We were talking about your growth potential in terms of the Bitcoin mining, and you said, 'Boys, all we need to do is plug in these machines.' Fast forward now to the end of the year, IREN has put up a phenomenal amount of growth in terms of data center capacity, exahash capacity, Bitcoin mining fleet efficiency. So the first question I wanted to ask before we get into earnings: how is this level of growth even possible, Dan? You guys are putting up more infrastructure, more machines than anyone else in the industry, and seemingly have it now just on rinse and repeat.
Oh yeah, I appreciate the compliment. I think it's a combination of a few things. Like, it comes from a founder-led business, if I'm honest. Like, Will and I, this is our life, and we're not corporate CEOs here. And there's no bureaucracy, there's no politics. It's just cutting straight through to what the right answer is. And we're wrong all the time, but it's the ability to recognize mistakes, pivot, and improve. And I think that process, that iterative process, has allowed us just to cut a lot of time and cost out of these processes and get to the right answers quicker. It's a mix of having the right people, process, and strategy. In Childress, we got 500 people on site every day helping to deliver on this. And look, it took a while to establish the right cadence, but you just keep persevering, you keep tweaking, you keep iterating, and we're now at the point where we're producing 50 megawatts of data centers a month, and it's largely copy-paste. It's just do the first slot onto the next slot, rinse and repeat. So looking forward to Sweetwater down the road, where basically the same processes, a lot of the same people will hopefully be able to help there.
Yeah, it was a great earnings call last night, and you covered a lot of information on there. And we encourage people to, you know, to go back and relisten to some of the points on there. We'll cover some of the stuff today, maybe take some of those questions in a different perspective. But I was really impressed. I mean, I said to Bryce before the podcast, we did our own podcast today, we covered a little bit of the earnings on there, and we talked about this last sort of three and a half years since your IPO and the fact of how the change in sort of retail investors, the popularity. You know, we go back to 2021, it was HIVE, it was Bitfarms, it was Marathon, it was Riot, those were the names that people were talking about. But in 2023, we saw this change, then we saw CleanSpark come through, we saw you guys come through, and 2024 really established you as probably the two most popular from the retail perspective. You know, and there's a lot of retail perspective out there in the land of X and all the other social media platforms, but YouTube really sort of set the tone for the industry, and both of you had massive growth strategies to achieve last year, and you both achieved those. Now it's easy to say when you talk through updates about, you know, energy pricing, price for Bitcoin, but I managed to go through the numbers this morning, and it looks like you did exactly what you were saying through all these updates, because it's never always the case. We always hear, you know, snippets of information come out, and then we go through earnings updates from mining companies, and it never sort of stacks up as the results that come out. But I had time to go through yours today and do some comparisons, and it was pretty amazing. The gross margin we looked at, Bitcoin mining, you know, 74, 75% gross margin in the mining business there, and producing a Bitcoin for 40,000, including all your sort of cash costs, so not just energy, but all your operating costs as well. And comparing that to sort of the other peer miner that released their results last week was CleanSpark, and they were sort of like in the 54, 55,000. So it gives you a 14,000 per Bitcoin advantage over your closest rival effectively in the market at the moment. So from your perspective, what else can you tell us about last night, very briefly, some of the key things that you talked through financially from the earnings update?
Yeah, and I think that cost per Bitcoin is based on a far lower average hash rate for last quarter, recognizing we only hit 31 at the end of the year. So I would expect to see that significantly improve again for this quarter. Oh, look, it's kind of you to say. It has been an interesting ride over the last three years, but as we say in Australia, head down, bum up, just focus on the process, make good decisions, focus on the fundamentals. And to be frank, like, public markets has been an eye-opener. Like, I've never had ambitions to be the CEO of a public company. I'm not sure I still do, because it's a very interesting environment where you've got a lot of different factors at play. But at its heart, you know, Will and I have got really large stakes in this business. It's our baby, we started it, we're focused on getting the right outcome over the long term. And obviously, when you start to deliver, the market's starting to recognize that. And we do appreciate all the support, you know, particularly of retail over the last few years.
Yeah, and you're quite right. We also discussed earlier, Bryce and myself on the podcast, the fact that as we go into this quarter, and we're only literally, what, about just over six weeks from the end of your quarter three, the first quarter of 2025. It never ends, it never ends. But actually, you're right, the fact that you've got to that hash rate and you're now operating 15 joules per terahash, because if you look at the, I think if I remember right from your presentation last night, your average hash rate for the quarter was about 22, and you're growing towards 15. So you weren't necessarily at that fleet efficiency for the whole period. Now you're at 15, and you've got the 31 exahash, so you're going to be producing more Bitcoin in this quarter, irrespective of where the difficulty goes. It would have to significantly increase to bring the amount of Bitcoin down in this quarter than you did in the last quarter. So you can have more revenue, more margin, and hopefully your operating costs, which I tend to class as generally fixed, really, you know, to a certain extent, unless companies are significantly growing quicker. What would that be sort of like your take on the next quarter?
Yeah, I think we've previously guided towards around $30,000 as an all-in cost. So $40,000 at the 22, if that's the right number you quoted, I would expect that to trend down significantly this quarter because, as you say, we've got 50% more revenue, but the costs haven't scaled. Spread over more, yeah, like, yeah, corporate overheads. This is the whole operating leverage of scale, and we've hit it now. The big ones are like RECs. So I think we're still the only Bitcoin miner that actually uses renewable energy because no one else buys the renewable energy certificates. And that's something we'll continue to do because we believe it's the right thing. You can't say you're using renewables if you don't have the certificates. And then it's things like property taxes, insurance, etc., that all just go line-item with the site. But obviously the people, we don't need a third CEO, clearly, and you don't need more CFOs. Yes, we're investing in growth, development, the AI side of the business, but those costs are relatively small in the context of the revenue line that we're now generating and will be generating, particularly at 50 exahash.
I mean, looking at market capitalization this morning, I mean, there was literally a hundred million between you and CleanSpark. So obviously the market, you know, is starting to give, you know, more credibility to you guys. We've seen that the share price and the market capitalization grow quite nicely over this last sort of eight, nine months, and, you know, getting closer and closer to the share price reaching a high. So do you think any of the sort of like the start of the HPC is starting to get built in, or do you think once maybe announcements in the future for any HPC, any of the sites will add to that?
It's hard to know, isn't it? And I don't track other companies' market caps that closely day to day. But that's an interesting point. Look, I just know based on conversations with our investors what they're thinking about. Look, we've tried to be pretty realistic about AI. Yes, we've got our AI cloud business, which has been growing, collocated alongside the Bitcoin mining machines in the same data centers. Yes, we've now announced Horizon 1, which I'm sure we'll get to, the 75-megawatt liquid-cooled. Yes, we're talking about strategy to unlock a potential 2-gigawatt data center hub in West Texas. How much is priced in? It's really hard to say. At the end of the day, we'll just keep being realistic about the process we're going through and where we want to get to. But a lot of these things, Bitcoin mining, we can control. I can give you a target. We're controlling our own destiny. Yes, there's supply chains that need to be managed, and we've done a pretty good job of that previously. But when you're dealing with third parties and relying on commercial agreements with customers, I actually feel it's a bit dangerous to try and preempt certain dates and terms because you're cutting across a natural negotiating dynamic with these counterparties.
Yeah, just staying on mining. I mean, I noticed you recently increased your hash rate to 57, but now yesterday you pulled that back now to 52, with obviously your target by the end of the year to get to 50. Is the extra two just filling capacity that you had available at Childress, maybe swapping machines over for more efficient miners, or just increasing with the megawatts available?
Yeah, I hate giving up hash rate. So going down from 57 to 52 hurt a bit. But when you're swapping it for 75 megawatts of next-generation liquid-cooled data center, kind of feels pretty good. Basically, that's utilizing the 25 megawatts in Block 6. So 75 of that 100 megawatts will go towards Horizon 1, and the 25 megawatts will allow us to get approximately two more exahash.
Yeah, yeah. I just like to say as well that, you know, with regards to the monthly update, where we get a vast amount of information. I mean, remember, I think, remember when we spoke over three years ago, when we started this place together, covering, you know, doing the analysis month by month, and it was literally a few sentences. And now it's like you've got to spend 15, 20 minutes just reading to go through all the information that's packed in there. I like the new look of your update. The photograph of Childress looks something, it doesn't look like a Bitcoin mining site to me. It looks something far more superior than that. And I noticed you mentioned last night, we don't build things in the desert in tin cans. Well, so it's great to get that information. I always tell the retail market, you know, manage your expectations because we've noticed a couple of miners this month have stopped producing monthly updates. And, you know, we're grateful that, you know, we've had this now for quite a few years, and we're hoping obviously it continues in the Bitcoin mining space. And also you very much include everything about your HPC setup as well. So we, you know, those people that have got a thirst for information feel that they've, you know, fully met their requirements when they finished your update. So again, it's great to see the new look of the update. Really impressed with the photographs, and hopefully that will continue.
Yeah, absolutely. I mean, we've got nothing to hide. So the more we can show, the better. And like, we've always viewed this as a bit of an educational exercise as well. We're on a journey together in new sectors, whether it's Bitcoin mining, which is starting to mature, whether it's AI. The more we can share, the more we can help investors understand how we're seeing the world and what we're doing. I think that's just going to help us over time.
Phenomenal. Yeah. Now we're going to get into the HPC AI component of the discussion next, which I think a lot of people are excited about. One final question I had from the earnings, you guys cited an energy cost in the $21,400 range. Can that be assumed that will continue at that level or perhaps decrease as additional power generation comes on? I know you just mentioned your all-in cost per Bitcoin you are guiding towards $30,000. So I would imagine the power component of that would at least stay consistent in 2025. But any commentary you can provide there?
Yeah, it's a function of a few things, right? So we've got a power cost of three cents a kilowatt hour at Childress since moving to spot pricing. Will it stay there? Will it be slightly higher? Will it be slightly lower? There's some level of risk there. The fleet efficiency obviously won't change materially. We're at the 15 joules per terahash, although there will be some natural improvement to account for the step up from the average hash rate of 22 to 31 at quarter end. What does the global hash rate do? Does difficulty increase? Sure, over time. So there's a few moving parts there, but directionally, I think it's likely to be close to that mark.
Great. Now moving into the next component here, the construction, the client updates on the HPC side of things. We noticed last night, and you just made reference to this, Bryce, you've decided to reallocate some of those megawatts into the Horizon 1 project. So 75 megawatts that were formerly or potentially going towards Bitcoin. We've talked to a lot of companies in the space, obviously, Dan, and a lot of them, similar to what you said on the call last night, are using kind of a pilot or trial approach to HPC. However, many of the other people doing this are on a much smaller scale. So our question was, this 75 megawatts, it seems like a lot to be allocated to a pilot project or a test project. How can you be so sure of this without having a customer in place, or how can you be sure that a customer will be available for this service?
Yeah, I think there's potentially a big difference here. We've got over 500 megawatts of test projects. Like, we've got that in operating data centers. And in those data centers, we're already operating an AI cloud service, servicing AI startups, scale-ups, increasing enterprise demand. So this is just a further evolution of that trajectory that we're on. And yes, we're using a slightly different cooling system with liquid and enhancing some redundancy with generators, etc. But it's not a material change from a business that was set up on the foundation of pursuing AI. And we often forget that. And we had that slide from February 2020 in our earnings deck. Like, this was always anticipated. It was build out a data center platform, bootstrap it with Bitcoin mining, be ready to leverage into higher-value use cases if they emerge over time. And we quoted AI, we quoted machine learning. So yes, it's playing out. We've got the systems, processes, the basic tenets of a data center there for multi-tenant capability. So to continue to evolve that makes complete sense. So I think the real question, therefore, is not on the technical side. I feel pretty comfortable there. It's more, is there a market for this and can we sell the capacity? And obviously, there's a risk we don't have a customer now. So by definition, there's a risk that we don't get one. But I think when you step back and you see the scarcity of liquid-cooled data center capacity in the market, all the Blackwells starting to ship, particularly in the second half of this year, and the inability of the real world to scale, you know, traditional data center operators, other companies, it's not easy to pull together 75 megawatts of liquid-cooled data center capacity like that. And in relative terms, our ability to do it this calendar year seems to us to be a real competitive advantage.
In terms of allocating that amount of power to Horizon 1, I mean, you just mentioned the fact that you're already operating GPUs in the same facilities as probably S21 Pro miners. Is there an opportunity, should things not turn out from a HPC perspective, to actually put S21s in? Or once you've made the decision now, you're basically forced with that cooling system, forced to using for HPC?
Ah, to be frank, I think it's almost unfathomable that it would come to that. But yes, like, the guys would find a way, and maybe the hydro models, for example. But also is the case that these data centers, the base build, the power shells, if you like, in which we operate Bitcoin mining, are only $650,000 a megawatt. So you could almost build a separate building next door and temporarily run miners, and they pay off that building pretty quickly.
Very quickly. Yeah. In terms of, I mean, in terms of the size of it, I mean, you know, I'm assuming you're looking for a host client to come in with their GPUs because the cost of you funding that would be significant. I mean, it's great that, you know, companies can do it on a small scale. It's expensive. You've already mentioned, you know, in terms of how many millions per megawatt. But when you're filling 50 megawatts with GPUs, that's a serious amount of capital required. And you're also looking at Sweetwater as well, one to, you know, to start developing in sort of like conjunction with what you're doing at the Childress site. So would that be fair to say?
Yeah, absolutely. I think there's a slight nuance here. At the end of the day, capital is next to unlimited if you've got the right business, right line, the right revenue line. So I won't rule anything out. For example, someone comes along, signs a three, four-year long-term contract for GPUs from a creditworthy counterparty, all of a sudden you're going to be able to secure a lot of GPU financing against that, i.e., debt capital, because the contract is there, it's bankable. So I wouldn't rule that out. Absolutely, we should stay flexible and open to those scenarios. But I am agreeing with you in the sense of the base case is just simple colocation, bring your own GPUs and plug them into our facility.
Yeah, I mean, we've only seen at the moment, we've only seen, I think, the two big contracts, one awarded with Core Scientific and Terawulf announced something more recently on a similar scale to what you've got at Horizon, actually. And we're hearing, obviously, there's going to be other announcements imminent. You know, the fact that I think something has been released with regards to Hut 8, you know, even prior to the new year, and that's not been formally announced yet, but every other announcement, it seems to be happening. Are you, this sort of backs up what you're saying about the, there's just, there's a requirement for power out there, and those companies who've got access to power and access to the fiber and the water that's required for HPC sit themselves in a very good position.
I think that's right, particularly coming into the new year. I mean, there's a different type of energy and intensity post 1 January, particularly from these hyperscalers, where I don't know what it is. Maybe it's people realizing that this power is actually scarce. Maybe it's people realizing that they are more short on power and DC capacity than what they thought late last year. Maybe it's just people re-energized after a long 2024 and coming back and ready to actually do work. But however you want to justify it, there's definitely an energy in the sector, and that doesn't surprise me to hear that other companies are also getting traction, and that's fantastic.
Do you think there's, I mean, it sounds, I mean, is there enough, do you feel you're in competition with other miners, or do you think there's actually so much out there that we're going to manage our own position going forward? I mean, eight months ago, we were sitting, of course, with Core Scientific, and the management team wouldn't take any questions about HPC. Now there's like a 180-degree turnaround. They're looking like they're going to use Core Scientific, like you're using Horizon 1 at Childress, and use that 600 megawatts on a site that's been effectively developed and just needs the buildings put up now to start delivering it. Are they competitors, or is it a case of actually there's so much in the space, you know, we'll do our plan with what we've got at Sweetwater 1 and 2, we've got Horizon, and, you know, can be done in fairly quick time because of the fact you've already sort of like done a lot of work at Childress.
Oh, if you look at that earnings presentation from yesterday, we've got a slide in there showing a 36-gigawatt shortfall in energy over the next three years. We've only got 2.3. I feel a bit of FOMO, like we're not doing enough. It's not enough. So you kind of look around, there's going to be a lot of winners. There's going to be a lot of opportunity for different people. We'll keep running our race. I feel like we do have somewhat of a competitive advantage with our ability to execute, the people, the processes internally, our track record of execution. But at the end of the day, like, it's a big market. There's going to be lots of winners, and that's great.
Final question on sort of on the Sweetwater is you have two, you've got two sort of sites that you can develop. Would it ever be an option to utilize one of those sites to sell and grow and use that funding to grow the main site then, or is that actually, it's not even an option? We know the future, we know there's growth and we can attract more return on investment by developing both of those sites.
