Hey, miners. Welcome or welcome back to the channel, McNally Money, the official home of Power Mining Analysis. Anthony and I could not be more excited. In today's video, we've got Adam Sullivan joining us once again on the program. He's the CEO of Core Scientific, and what a phenomenal 2024 it was for this company. Not to mention, they're kicking off their monstrous HPC contract with CoreWeave in Q1 of this year. We've got so much to talk about. Before we get into it, take a second, smash the like button, you guys. Big help to myself and the channel, especially in getting this content out to other people like you who may find value.
If you're not already subscribed, McNally Money, feel free to join and let us know in the comment section below if you managed to catch the CORZ trade in 2024, if you're currently holding shares, and your outlook for 2025. Now, with that being said, let's get into today's interview. Okay, guys, so that's right. Huge interview today. Adam Sullivan, CEO of Core Scientific, back by popular demand. Adam, I was just saying when we were putting together our question list, we were inundated with questions for Core Scientific, an extremely exciting quarter, year, and coming off 2024, which was a tremendous transformational year for Core Scientific. So, so much to talk about. Before we get into it, though, thank you for making the time. Great to see you again.
Thanks for having me on. Happy New Year, everyone.
Most definitely. Happy New Year to you too. So, I had just mentioned 2024 was quite the year for Core Scientific. I think anyone in the sector would agree. Do you mind walking us through, Adam, just a few of the top highlights in your mind?
Yeah. I mean, it was a quarter-by-quarter exciting year for Core Scientific. You know, Q1, we started January 1st. We were still in Chapter 11. We went public on January 24th, and then shortly thereafter, we announced our first CoreWeave deal of 16 MW in Austin, Texas. You know, we had a kind of a blockbuster Q1 in terms of our Bitcoin mining business and the numbers of Bitcoins that we mined. And then Q2, we were able to announce our first conversion deal with CoreWeave, so a 200 MW deal, January or June 3rd. Shortly thereafter, in the beginning of Q3, we actually converted. We were able to convert the mandatory convertible notes based on share price performance.
Then we went on and raised and restructured our capital structure, eliminating all of our previous Chapter 11 debt, putting some cash on the balance sheet, finished out the remaining 300 MW of options with CoreWeave, and then announced another convertible note deal. And so now we sit here today in January of 2025 with a significant cash balance. We signed one of the largest hosting contracts in data center history and potentially one of the most profitable hosting contracts. And so we feel very, very good about entering 2025 with a really strong platform that we built in 2024.
No, that's great. That's great to hear, Adam. I mean, 2024 was a phenomenal year, but we're coming to 2025 now. So really, you know, what more can we expect this year to whet the appetite of the retail investment and also those institutional investors that seem to be piling into Core Scientific on a weekly basis?
Yeah, we really set up this platform to build on in 2025, and we laid out our goals through 2027 to build over one gigawatt of operational infrastructure for HPC. In 2024, we were able to set all the groundwork. We built out our team, we built out our asset base. Entering 2025, we have a significant portfolio that we can continue to sell and market, and we're going to continue to maintain a very strong and profitable Bitcoin mining business. You know, we've announced that we have 400 MW dedicated to Bitcoin mining. We have those 15 exahash Block chips coming in in the second half of 2025. We expect to have one of the largest HPC businesses that we're going to be building out over the next three years. We'll also maintain a very strong Bitcoin mining business.
Yeah, that's awesome. In terms of the HPC hosting, because obviously that's a key part of this discussion today, with regards to the contract itself, can you walk us through in terms of the sites that are now allocated for HPC? I mean, you've got to go onto your website. There's a whole raft of, you know, locations of all the sites, but, you know, this is a big contract with CoreWeave, so it takes up more than one or two sites. So you can talk us through that and also sort of put a bit more color on sort of like the refurbishment program, what that entails. I mean, you know, is it a case? I mean, I heard the word demolition mentioned this week in an update from Core Scientific, and I was surprised about that there. That seems you're going right down to the basics and rebuilding.
If you can put some color onto that, it'd be really helpful.
Yeah, absolutely. So we've said that there's five sites being converted for this contract that we signed with CoreWeave. We've announced so far Denton, which is going to be one of the largest supercomputers in the United States, which is incredibly exciting. We also had the groundbreaking ceremony in Q4 as well for Muskogee, Oklahoma, which are two fantastic sites in our portfolio. Now, we have three other sites that are going through conversion processes. We've said previously that, you know, we don't like to announce the sites where we're going because, you know, that means we're still in discussions on whether it be more power approvals, what the power price might be for additional power allocations. So these are things that, you know, we're still working through on some of these new or, I should say, expanded power contracts that we're working through.
It's providing us an opportunity where we can continue to work with the local community, work through incentives, work through utility contracts, and put us in a position where when we do announce it, we know we've gotten and extracted as much potential power as possible from those sites. That's happening for three more sites. If we shift to talking about conversions of existing sites, you know, for certain sites, we are converting the buildings. In other situations, though, we are going through a process where either we're removing an existing facility and building from the ground up, and that is usually based on whether it be cost or whether it be speed to deployment. This is a significant amount of capital.
You know, it's anywhere from, let's call it, you know, $5 million-$8 million per megawatt, depending on the site, depending on what we're building out. You know, Muskogee, Oklahoma is a great example. That's a greenfield site outside of the fact that we already leveled the ground. We already had the ground leveled. We already had the high voltage substation in place, but everything else essentially has to be developed there. So that comes at the really higher end of the price range. But, you know, we feel very strongly about, you know, the fact that we've secured the entire supply chain for 2025 for all of those developments and then securing the remainder of the items that are left over for the 2026 developments, which, if you recall, we've said by the end of 2026 or by the end of 2025, we'll have 270 MW fully operational.
By the end of 2026, we'll have the full 500 MW that is currently contracted with CoreWeave operational.
In terms of, you've mentioned about cost per megawatt there, and that's a significant amount. That's not something to, you know, to attract. The funding that you're receiving from CoreWeave itself, they're going to give you an advance amount of funding. How much of the sort of refurb will that actually cover in terms of the HPC contract?
Yeah, so our contract with CoreWeave is they cover 100% of the CapEx. And so if they decide to change based on whether it be their requirements or their end customer's requirements, whether it be redundancy, whether it be, you know, the style of building that they would like to have, any of those increases in costs are still passed through directly to CoreWeave. So we have no burden. The only burden that we do have is paying them back $1.5 million per megawatt. And essentially that happens over the course of the first about two years, just over two years at a 50% revenue credit against their contract.
In terms of when, I'm assuming those payments are starting effectively now because you're doing the refurbishment now. In fact, if you're doing demolitions, that's all part of the capital requirement. When do you anticipate sort of like revenue coming in from the first phases of the contract itself?
Yeah, so we've said that we expect to have 200 MW by the end of Q2. And so that's going to come in a ramp schedule over the course of 2025, not only the first half, but also the second half to get to that full 270. So those are going to ramp on kind of a building by building basis. And so, you know, I would say we would expect to see revenue over the course of the first six months of this year as we work to build towards that 200 MW.
The reason I bring that up, Adam, is I'm still scratching my head at why the market aren't sort of like fully factoring all this in. I mean, we've seen the numbers, we've seen the table. I produced a table using your numbers, and, you know, it really is, you know, a phenomenal contract to have, but I just don't feel, having spoke to three analysts in Sweden at the end of last year on a trip to a site over there, we talked about the Core Scientific. These are Wall Street guys that you know very well. And I said, can you give me an idea of how we should value these companies? And they said, even if you were prudent, you'd use 15 times earnings.
I'm scratching my head really, even in those early years, because you still get a, you know, a percentage of the earnings, even though you're paying back an element, you'll still get revenues starting to come through. And yes, from sort of like year three to four, you're going to start getting, you know, significant amounts. I mean, well over $500 million a year of net income. And I keep repeating on the podcast, it's net income. It's not, this is not more cost to come out of it. This is net. This is something that's unusual, even in the Bitcoin mining space, to have companies come out with net income. We go back to last quarter, only one miner in North America had a net income. And that was because of a couple of one-off transactions that really helped them in that period.
Everyone else was red and the whole lot should have been red, but one was green. This will be nice in 2025 when we start seeing that green turn for pretty much most of the miners in terms of self-mining, but especially you guys in terms of the HPC hosting. Now, we know about the CoreWeave deal. We know you brought out updates about your, you've got more power because I think you've allocated 900 MW in total. Now, thinking about what you've got CoreWeave plus the ancillary, that's probably over 700 MW there, but that still gives you about 130 MW-150 MW of additional. What are the conversations that you're having now?
And bear in mind, we were there in June at the Denton facility when we were listening to the first part of that contract of 200 MW, and you indicated that the phone was hot and there was plenty of conversations. Are you still having those type of conversations now, knowing that you've got a significant amount of power that you're putting towards, you know, extra deals in this space?
Absolutely. I mean, the market demand right now is still very hot. I think what we can expect to see over the course of 2025 is a lot of more deals being announced, especially the larger deals. I think over the next 12 months in 2025, we're going to see all of these large mega site deals get announced. Not necessarily, they're not going to be built in 2025. They may be built over the next three to five years, but I think this is the year where all the hyperscalers lock down all of their large scale sites.
We're working closely with a number of different not only hyperscalers, but also large enterprises on securing a new contract with one of them because we feel like we have an opportunity to potentially secure multiple contracts over the course of 2025, which will put us in a position to achieve that one gigawatt fully operational by the end of 2027. You know, our goals are bigger as you look out the next five to 10 years, but over the next three years, if we achieve this one gigawatt of fully operational capacity, that'll make us one of the largest data center companies in North America. If you look at what kind of our investment thesis is now for investing in Core Scientific, we signed a highly profitable 500 MW deal with CoreWeave.
That's not only one of the largest, but potentially most profitable contracts in data center history. So we have a great platform to start on top of. And then our growth story, I mean, we're going to be growing faster than really any other data center company on a percentage basis over the course of the next five years. So you take a company with better margins, better growth profile, that puts us in a position where, you know, we think we should be trading at a premium multiple. And, you know, we believe we can execute on this given the team that we've been able to put together and the asset base that we have today.
You talk about team and you raise a good point because my next question is about obviously having all these new sites coming on the refurbishment. It sounds like you're having to put together a bigger team to take on this new business that you're going to be providing with CoreWeave. How's that recruitment process been going? And, you know, have you had a lot of interest in the sites that you're already utilizing for this business?
Yeah, I mean, talk about the team. You know, the team that we've been able to assemble that we had when I joined this company that we've been able to add to over the course of the past, you know, six to twelve months. I mean, we are highly rigorous in terms of the talent that we recruit. You know, we're looking for top-tier talent and we've been able to be in a position where people want to join this company. You know, you take top-tier talent from any of the other large data center companies and you put them in a position where they can be accretive from day one and add value to the company. And on top of that, they know that, you know, they're receiving equity as part of joining Core Scientific.
The fact that they can directly influence how that equity performs puts them in a position where they want to come to Core Scientific. You know, the 500 MW asset base, obviously incredibly interesting to not only engineers, but also construction people as we're building out infrastructure that's really being developed in some cases for the first time. We're developing infrastructure for GB200s, which haven't been really deployed at scale before. This is providing all these people who have been building traditional data centers for decades in a position where they can build out the next generation of data center. People are excited. You know, we've been able to assemble quite the team. You know, we're taking very senior level people from the other large data center companies and bringing them to Core Scientific.
You look around the table today. I mean, we're surrounded by a lot of folks that have been in this industry for decades. You know, it feels great when we walk into meetings with many of these potential clients and they know everyone in the room. They're like, "Oh, we built a data center together in Georgia. Remember that? Remember that time?" You know, it's like those are all the fun conversations that you get to hear about data center development over the course of the past 30 years.
We were extremely fortunate to be there in June and met a great deal of the senior management team there. And we can talk firsthand of that team. And I mentioned the team in many of the podcasts. It's the Core Scientific team that delivered where we are today. My last question on the sort of the HPC is obviously we're now seeing a number of other miners. One's already announced. We've got TeraWulf to announce their contract recently. And it looks like we may have another miner imminently about to announce as well. They've got some planning decisions going through this week. So maybe we get an update from Hut 8 on how that's going along.
But in terms of, you know, and I don't want to speak, you know, too much about it, but in terms of what you're seeing at the moment, and you were the first one to come out with the really big contracts, you're laying everything out there and you're now seeing other companies contract with other clients. How do you feel about your contract in terms of the numbers and looking at what other competitors are doing at the moment? Do you feel you've got a great deal or do you feel actually it's a market with potential to get even better?
Yeah, first I want to say congratulations to Paul and the entire TeraWulf team. It was great to see another contract. You know, we were really excited when we saw that across the wire. You know, as we look at, you know, our contract versus not only the TeraWulf contract, but, you know, the rumblings that we're hearing in the market, you know, the CoreWeave contract that we signed was truly exceptional and best in class and one of a kind. You know, obviously we're looking to continue to expand with CoreWeave at some of the existing sites where we're getting some additional power. And so, you know, looking at our contract versus where the market stands, you know, we're still extremely, extremely confident that this is potentially going to go down as the best data center contract in history.
You know, as we think about continuing to negotiate contracts in the future, you know, we've said there's going to be a mix of, you know, whether it be us paying for a portion of CapEx or potentially even 100% of CapEx, whether it be through debt, a combination of debt and equity. You know, we're seeing rates though in a very strong position, referring to lease rates here, and so, you know, we're excited about where we're going to sit by the end of 2025 with potentially more contracts being signed, and I think, you know, we're still in a position today where we can sign better contracts than the rest of the market because people have a significant amount of trust in our team, and I believe we have the right asset base to deliver high quality infrastructure to clients, hyperscalers and large enterprises.
That's awesome. We'll touch on a bit more about funding in the second half. I know it's Bryce, probably got some questions on the self-mining part of the business.
Oh, I certainly do, Anthony. Yeah, if you've noticed, viewers, we've split it up into sections here. So now we're going to talk about self-mining and HODL, question mark, potentially. So self-mining, interesting. Just on the HPC, Anthony, you had talked about that as the deal of the year last year. Then we started calling it the deal of the decade. Now it's sounding like the deal period in HPC AI history. So it just continues to get better. But you're right, Adam, as we compare to other deals, a lot more transparency and it looks like a lot better numbers from what we can see with the CoreWeave situation. So again, congratulations there. On the self-mining, we brought this up on today's podcast as well. There's been so much excitement around Core Scientific and CoreWeave about the HPC, the megawatts, the retrofit.
People often forget you guys are the most consistent, reliable, and largest net producer of Bitcoin that is in-house mined, period, as well. So you get both elements of this company. Speaking of self-mining, we know you have the big order with Block, Inc. Can you talk to us about when we can expect to see those installed and how that's going to impact the fleet?
Yeah, yeah. So we've said publicly that we expect to receive those over the course of the second half of 2025. So whether it be a portion or the entirety of that 15 exahash to be installed, you know, a majority of that is actually going to be refreshing existing machines that we have on our sites. You know, if you look back at what our mix of machines is, we do still have a significant amount of not only traditional S19s and an S19 Pro. So this is going to put us in a position where we'll be able to increase our terahash while maintaining the same amount of exposure in terms of megawatts to Bitcoin mining. You know, that's a business that, you know, we're looking forward to being able to refresh the machines.
You know, some of that fleet is aging as we look at, you know, them being operational for, you know, three plus years. And, you know, at the point in time now, it's where things start to break on machines. You know, you're going through more significant repairs. And so this 15 exa hash machines is going to be fantastic for our portfolio. We don't know, we haven't necessarily said where we're going to end up in terms of hash rate expectations for end of year. But, you know, I think a lot of that's honestly going to be dependent on where hash price is because the Block chips are going to be able to help us provide so much more flexibility in terms of not only total terahash exposure, but also efficiency of machines on a joule per terahash basis.
And so, you know, we could end up, you know, with, if hash price is performing well with, you know, significantly more exahash online than we were expecting today.
Yeah, that's really exciting. Another comment we talked about today was just the consistency you've been able to maintain with the Bitcoin division despite the moving, the retrofitting, the buildout for the HPC. So really incredible there. You touched on an interesting point, some of the aging fleet that you're upgrading, Adam. As the price of Bitcoin continues to increase, we see the economics improve. Is there a point at which maybe you could remain with some of these older machines as part of your portfolio?
Yeah, it's not necessarily profitability that is the issue. It really comes down just to the degradation over time of those machines. And so as they degrade, you know, you're putting more into maintenance of maintaining that terahash. And so I think, you know, as we look forward to the 15 exa hash of Block chips, I mean, that's going to be highly profitable and highly accretive to our business that we're excited to be able to do that across the two remaining sites we have for dedicated Bitcoin mining.
Great. And I just wanted to clarify. So you're saying as hash price and Bitcoin price improves, you haven't committed to a specific exahash target, but you're going to monitor that and you've got the chips and the machines as needed to scale.
Absolutely right.
Okay, great. That was a specific subscriber question. So I just wanted to make sure we hit that one. Now you've talked about the allocation between HPC and Bitcoin mining, about 400 MW right now allocated to Bitcoin mining. We've talked about some of the site opportunities expansion in the United States. Are you still looking for Bitcoin mining sites at this point?
I would say we're definitely still looking for more sites, but not necessarily on the Bitcoin mining side. You know, our focus today is on the HPC side. The 400 MW we have for Bitcoin mining, those are fantastic sites. We have great power contracts. We're able to participate in power programs there. That puts us in a really strong position to maintain a profitable Bitcoin mining business, let alone that infrastructure that's on those sites will allow us to introduce those Block chips and have them perform exceptionally well. So we're happy with the 400 MW, but on the HPC side, we're still looking at a number of different acquisitions on the site side, including potentially even portfolio sites. So that's where our focus is, and it's really moved away from looking at new sites for Bitcoin mining.
Makes sense. And appreciate the transparency. Now, another thing we're looking for some transparency on. I've been tweeting this out over the last few weeks. One of the keyboard warriors out there, Adam, has seemed to have found a Core Scientific HODL account or address. We've been watching it grow and grow. I think it was around $27 million or something in that neighborhood. So Anthony and I on today's podcast, when we talked about your production results for December, we promised we would ask the subscribers, what is the HODL strategy? Do you have a HODL? And how is that going to look in 2025?
Yeah, we're looking at maintaining, you know, a certain portion of our balance sheet to be held in Bitcoin, not only for exposure, but also for potentially using in hedging as well as we look forward towards, you know, potentially taking advantage of opportunities where hash price increases significantly over a short period of time. But, you know, as noted in our production report that we released this morning, we did not sell all of the Bitcoin that we mined over the course of December. And that's a strategy that we're looking to continue to implement over the course of 2025.
Any chance we'll be able to get visibility on those production reports, Adam, similar to some of the other miners?
I mean, we're just going to continue to be announcing in terms of the Bitcoins that are sold at the end of each month in our production reports. And so that's really the form of communication that we'll be utilizing to communicate to the market.
That's okay. We'll keep you busy, Anthony, with that one then.
Who's that guy with the wallet in the back? But moving on to sort of like funding now, we've covered a couple of topics in funding already in terms of what CoreWeave are going to be helping to pay in terms of the refurbishment and in terms of the advanced payments, et cetera, et cetera. But one real big story that came out towards the end of last year was the successful convertible note, the $550 million convertible note. Can you tell the subscribers, listeners, what's the plans for that at the moment?
Yeah, so the first thing I'll say is we're not using it to buy Bitcoin. We have really three outlined purposes that we're going to use it for. The first is in our negotiations with these large hyperscalers and large enterprises, we had to have enough cash on balance sheet because you do have to pay for, you know, if it's 20%-30% of a total cost of a buildout, they have to pay for it in an equity check. We needed to have that cash on balance sheet to be able to negotiate with these potential counterparties and some of these larger deals that we're working on. The second is looking at new site opportunities. So we need cash to be able to go out and continue to be active on new site acquisitions.
And then third is looking at pre-buying parts of the supply chain for some of these sites that we're acquiring or that we have acquired or in the process of acquiring. And so, you know, that's something where contract negotiation or negotiation really on engineering, not just the contract details, you know, that can take months. And so we don't want to be in a position where we lose and have timelines slip because we didn't purchase the long lead items ahead of time. And so we're looking at potentially pre-buying some of these long lead items for some of these new sites so that we can be in a position where we can negotiate with a counterparty and continue to maintain a firm delivery date with those clients. And so, you know, that cash is going to great use.
We're looking forward to being able to deploy, you know, a portion of it over the course of 2025.
Yeah. And I wasn't implying about whether to buy Bitcoin, but you make a good point. I mean, in terms of being a success, and we've seen other North American miners go down the same route. We've seen CleanSpark follow your line in terms of using it to grow the business. And we've seen the likes of Riot and Marathon. Marathon done two in the space of literally two months, but they purposely bought Bitcoin with theirs. Do you feel it's another lever that you can access in the future with the markets being the way they are and the fact that you've got this amazing deal that stands up for potential investors in convertible notes?
Yeah, I think what we're seeing right now in the market is that they're having a hard time continuing to digest all of this paper just to buy Bitcoin. I think the market and investors are getting relatively full of that paper. Otherwise, I think we would have seen MicroStrategy come back and hit another convertible note. You know, they pivoted and announced a different type of capital markets transaction. But I think the market, you know, looking back over the course of, you know, Q4, digested billions of dollars in paper just to buy Bitcoin, and I think we have a much different story here. I think our equity story is relying on a thesis that we're going to be developing one of the largest data center companies in North America.
I think as we think about capital market strategy over the course of 2025, I think it's going to come down to, you know, what's going on broadly in the market and what type of opportunities arise for us. You know, it's hard to forecast whether it's going to be converts if that's the right tool, whether it's traditional debt, whether it's, you know, broad just traditional equity capital markets transactions. We don't know yet because we don't know what the opportunity is going to be. We don't know where the market's going to be. And it's just really hard to forecast the future.
I mean, obviously, I've got a couple of questions. In terms of funding what you've got at the moment, and you've got this convertible note, you're obviously bringing the Bitcoin mining revenue. You're covering your costs there. You're hodling a little bit of the production. What sorts of, you know, big additional costs? I mean, in terms of the Block chips, are they covered already or is it a case of we've still got to find additional funding in the future to cover the elements that we've got in the strategy as we know it at the moment?
For the elements that we know today, we have enough cash. Now, if there's some large transaction that we may look to undertake, that could change the story, and so that may present an opportunity where we do need to go out and raise additional cash, but like I said, it's going to be so dependent on, you know, what's going on, what the opportunity is, what the return is on that type of investment, so it's so hard to know where we sit today, what's going to be happening in the future.
The good thing is you've got at least a few more levers to sort of like play with. And one of those levers, tell me if I'm wrong, I mean, you've come to the anniversary of leaving Chapter 11 back in, I think it was 23rd of January this year. With that anniversary, I'm assuming that you're eligible then to apply for an at-the-market shelf offering. And is that something that's one of those levers that you might look to pull in the future?
You know, so it's January 24th is the day that we are shelf eligible. Yeah. I would say it's really going to come down to market conditions and do we need more capital. What's the opportunity to raise or for us to deploy that capital? You know, we're looking at a number of opportunistic things today. You know, we're a company that is looking to continue to grow aggressively, and so we're looking at a number of different types of opportunities. From where we sit today, we haven't made a decision about which direction we're going in terms of whether it be any type of capital raise, first of all, but if there is a capital raise, what form that's going to take.
And finally, we've asked this question before, but in terms of the CORZ.W warrant, have you seen much progress on those being activated or is this going to be a long journey to think recovering that capital?
Yeah, not too much progress today. I wish there was more. There's still a significant amount of option premium embedded in that warrant. And so there's two things that can affect that: time decay, which every single day we're decaying a little bit more, which is going to reduce that option premium. But as stock price increases and as we kind of break through new handles, whether it be the $20 handle, $25 handle, those things are going to continue to push that option premium because the in-the-moneyness is going to be so high that that option premium will migrate towards zero. And so, you know, we're very hopeful that with some additional time decay, with some better stock price performance and pushing us into the 20s, you know, we'll be in a position where more of those warrants choose to exercise before the exercise date.
Yeah, great visibility into the warrants there, Adam, and we appreciate that update. Now, next topic here or section is really moving into the future of Bitcoin mining, the industry, obviously a lot of political developments and updates even here today in Canada, with that being said, what do you see as some of the biggest challenges facing this digital infrastructure or Bitcoin mining sector in the next year or two?
I think it's really going to come down to power generation. I think that's going to be one of the biggest headwinds to the continued development of this business. You know, we look at the past 10 years, really the past 30 years in the United States, we've been atrocious at building transmission lines from where power generation is to actually bring it to major metropolitan areas. You know, that presented a great opportunity for Bitcoin miners over the last four to five years, being able to find this stranded power across the United States. As we think about new sites and new opportunities and utilities that we're working with, you know, they're looking at also having us co-locate closer to power generation because of all this stranded power across the U.S.
But the major issue with that is latency, really the fiber latency back to a major metropolitan area, which is a major focus for everyone developing these next generation data centers. And so it's really about bridging that gap between finding sites that fit well based on where the actual generation is and proximity to that major metropolitan area with that fiber connection. So I think the big headwind here is being able to find sites that fit into that kind of perfect basket. And, you know, that's not always that easy, but I think we have great talents from this new administration. They're focused on new power generation. They're focused on Bitcoin being mined in America and their focus on artificial intelligence infrastructure also being located here in the U.S.
I'm hopeful that we see positive regulation for us to be able to continue to build out across the United States.
Just a quick question, because you mentioned power there and across the country. In terms of power cost, which varies significantly state to state, how much of that had an impact on the choices for sites that will remain self-mining? Because, you know, yes, you incur a power cost, but the margins for self-mining versus HPC are so drastically different that you can probably use higher cost power and not be impacted as much. So the question is, I mean, was that part of the process or was it a case of there were so many other factors, power price was just one of them?
Yeah, power price is really just one of them. Also comes down to latency back to a major metropolitan area. I would say in terms of how hyperscalers and other large data center developers think about power prices is, you know, they start to get sensitive around the eight- to nine-cent range in terms of when they start to feel the pressure. So, you know, all of our sites, you know, the great part is all of our sites were in locations that had favorable power costs. And so even moving to this full 24/7 power type contract, you know, doesn't bring them above that level where they feel like there's pain from the power cost. So, you know, we're excited about that, the fact that we had a full portfolio of sites that had that type of opportunity.
And then, on the Bitcoin mining side, you know, those are really the two remaining sites are two of the least expensive sites from a power cost perspective that we had in our portfolio. So, it puts us in a really strong position to maintain a profitable Bitcoin mining business.
And in terms of where we are at the moment, I mean, I switched on the news, I mean, I'm based in the UK, obviously, and the news in America at the moment is about the cold weather. What are you seeing with regards to the power costs now compared to where you've benefited this year? And believe you me, I looked at the Q3 numbers and you were one of the most profitable miners in terms of cash costs. You know, I look at Hut 8. So you have a really low power. Are you seeing a change in the power? We know you don't put your power costs on your updates, but one miner did this month and it was noticeably a massive increase, about 50% on what they paid in November.
Yeah, I mean, you have to remember a lot of our portfolio is in regulated utilities. So you don't see as much of that volatility as you might see if you were more heavily weighted towards ERCOT. Hard to say where we're going to fall for 2024 where we sit today. But, you know, we experienced favorable power costs over the course of 2024. So from the initial number that we put out in terms of guidance for power costs on the Bitcoin mining side in Q1 to where the next few quarters rolled out, you know, those were numbers that were revised lower and numbers that we beat. So, you know, we're excited about the favorable power costs in 2024.
Now, we were down in Denton, we saw your repair facility there, and you also gave us visibility into your software stack, your tech stack, Adam. Obviously, a lot of insight you can gather from these machines on the software technology side. Can you speak to us a little bit about that on the HPC AI side of things? You mentioned in a lot of cases building buildouts, you're one of the biggest HPC builders in the United States. What is that learning process like on that side of the business?
Yeah, I would say GPUs have downtime just like the ASICs for Bitcoin mining do. And so we're working on an overall strategy around maintenance and repairs of some of these GPUs. And this is something where we were one of the largest ASIC hosting firms in America. It puts us in a really great position to be able to be, you know, a leader of this repair and maintenance services as well on the GPU side, just because we have all that muscle memory from doing this on the ASIC side, developing specific tools for specific types of repairs. You know, these are all things that we are very good at doing. And it just shows the breadth and depth of our team that we can replicate this again on the GPU side. That's the one we're working on today.
Now, would that maintenance be a value add or an enhancement to the existing contract? As I understand it, you basically provide the building, the shell, they bring their own GPUs and units. Are you looking to maybe extend that into a maintenance and connection?
I would say we're looking at, you know, a number of products that can be offered to not only CoreWeave but also that may look to have us have a much more hands-on approach to the site and actually touching GPUs and maintaining the GPUs, so I would say there's a plethora of different options available to us in 2025, and we're looking to continue to expand our offering.
It should never be underestimated the amount of work that's required to keep a fleet operational. I was amazed at the numbers. I think we were there in June. It was like the six-month numbers on the screens in front of us. It sort of blew my mind as to how many, you know, some will probably obviously be very simple repairs and some, you know, maybe a bit more work required, but it should never be underestimated. We've had other, like the likes of Compass Mining on the podcast, and they do some in-house, but they do a lot remote as well. You know, rigs have to go away and come back. You guys have managed to, you know, set up repair centers so things can be put back to work, you know, in very short space of time and keep that hash rate moving.
That was really. That's really impressive there. In terms of HPC and artificial intelligence, do you see, you know, the AI machine learning impacting Core Scientific operations and the services itself?
You know, we utilize some artificial intelligence machine learning. We use it on the Bitcoin mining side today, which is, you know, as we evaluate how machines are performing, what firmware settings they're on based on all of the different variables we take in, whether it be weather, wind, ambient temperature, precipitation outside, humidity, all these different factors come into play. And we actually have our own firmware stack that actually helps adjust the entire fleet or on an individual machine basis based on whether a machine starts to experience some type of heat poisoning and starts to heat poison the machines around it to actively control for those types of outbreaks inside of a facility. And that's utilizing data from, you know, we've had over 700,000 machines in our facility. We've collected data on all of them.
It's put us in an opportunity where we are very good at maintaining all of the settings on these machines to maintain a higher uptime, and so we've used it in the past. We continue to deploy new services on the machine learning side, and we're looking at deploying more items internally on the AI side because it is truly, truly interesting what many of these things can do. Given that we're providing the infrastructure for so much of this development and learning capabilities that these machines are going through, it only makes sense for us to deploy a lot of that internally as well.
Just got me thinking, Bryce, is this the real Adam we're speaking to, or is this an AI virtual Adam?
There you go. Digital twin. Now, Adam, I think we have one each left, and then we'll let you go for your day. You're doing great. Final one, and this is relevant now more so than ever with the addition of the HODL, but you guys obviously have a lot of trade secrets, a lot of intellectual property. There's the whole cybersecurity side of things, and of course, protecting your digital assets themselves. Can you speak to us just quickly about how that all looks within Core Scientific as well?
You know, we have an exceptional information security team and physical security team. This is something that, you know, as an early Bitcoin mining company, you can't take too lightly. And so it's put us in a position where, you know, we have very strong physical and digital security that's put us in a position where, as we go through all of the audits, all of these different things that are required and due diligence that's required as part of working with hyperscalers, you know, they're impressed with the infrastructure that we've built out, how we protect our facilities, all of that security. And so, you know, without going into specific details, you know, I would say we're a very highly advanced company from that perspective, and we really pride ourselves on how strong our control environment is.
Final one for me is more along the lines of what metrics are you using to determine when we, I mentioned already in the conversation about hash price and hash cost for self-mining, but in terms of metrics using for HPC, are there any sort of like that standout metrics when you're doing compare and contrast as to through negotiations as to what you want to achieve?
Yeah, I'd say it really, it's really about where do we fall in the barbell of paying for all of the CapEx ourselves versus them paying for all of the CapEx. And those are the two ends of the barbell. Now, the rate decreases as the customer pays for more of the CapEx, but there are sweet spots and stopping points along that barbell that it might be them paying for a certain portion of the CapEx to do a small rate buy down. And so these are things that we're utilizing over the course of our negotiations.
And one of the main reasons why we did that last convertible note raise to put us in a position where we can go all the way to one side of a barbell where we're utilizing 70%-80% debt and 20%-30% of an equity check to get the highest rate we could possibly get because the ROI looks more attractive than them paying for a portion of that CapEx. So, you know, we feel very confident with the capital that we put on our balance sheet that we have that full spectrum to play with today, and it will put us in a stronger position as we negotiate these contracts moving forward.
In terms of final one, in terms of, you know, with the CoreWeave contract that is at the moment and you have this additional amount of megawatts available, can you concurrently build out other sites at the same time, or is it a case of it's timelines now? CoreWeave will take place over these years, and then when we've got a, you know, a chance to build out another contract, that will be, or can you sort of like run parallel projects at the same time?
Our team, not only the existing team, but as we continue to add to it, has additional capabilities for 2025 and 2026 on top of what's already been announced with CoreWeave. So, you know, we feel very confident in our execution capabilities getting to that one gigawatt. To get us to that number, it's really going to come down to booking those contracts and having an opportunity just to be able to prove that we can build that one gigawatt over the next three years.
Yeah, absolutely. Adam, it's great to hear, you know, everything's working well. You've got a great team. We've been privileged. We've seen the great team at Core Scientific, met most of the senior management team there. We were part of a great group assembled in there and every one of the people that were in that meeting back in June there, especially on the journey back to Dallas, was extremely positive. The positivity remains with Core Scientific. You're quite right. This is a deal of the century. I'm just waiting myself for the market to really, you know, accept that and sort of build that price, build that pricing that should be there, but you know, as our early podcast for 2025, I'm really pleased that it was you that came on, Adam, and fulfilled that. Thank you very much.
Thank you.
I certainly am too. We had 32 questions. We tried to poke a hole in the core business model, but even the simultaneously buildouts at the end, you said you're capable of doing, Adam. So very good podcast, you guys. Thank you so much for submitting the questions. Adam, we wanted to give you the opportunity just to close out any potential big catalysts we're unaware of or anything maybe we've missed in the podcast. But as Anthony said, Core Scientific, you guys, one we've been watching closely all throughout 2024, and 2025 promises to be even bigger.
Yeah, I would just say, you know, we had that third outstanding catalyst for 2024. That's something that we're working diligently on over the course of the first half of this year. And, you know, if you look at what we did over the course of 2024, you know, we built the platform that we are going to continue to build on top of in 2025 and beyond. And we couldn't be more excited about it with the team that we have, the assets that we have, the opportunity that's ahead of us is to become one of the largest data center companies in North America. We have high confidence in our ability to execute on that plan and be one of the largest companies in the data center space in not only North America, but potentially the world as well.
Phenomenal. Well, congratulations on a very successful year and awesome start to 2025. You guys, if you have any additional questions for Adam and team, leave them in the comment section below. Let us know if you're currently holding shares of CORZ, what you think about the HPC contract with CoreWeave, and of course, your outlook for 2025. Thanks so much. We'll see you back here.