Good morning, everyone. Thank you for joining H.C. Wainwright's twenty-sixth Annual Global Investment Conference. My name is Michael Donovan. I'm an equity research associate here at HCW, and today I have the pleasure of introducing Adam Sullivan, CEO of Core Scientific. Core has approximately twelve hundred MW of contracted power, five hundred devoted to Bitcoin mining and seven hundred devoted to HPC, of which five hundred can be directly sent to servers. Adam, the floor is yours.
Thank you so much. So Core Scientific is a leader in digital infrastructure. As was mentioned, we have one point two gigs of contracted power, and that has allowed us to develop and build one of the largest footprints for Bitcoin mining over the past seven years. We've mined more Bitcoin than any other public company in North America, and given the fact that we own our infrastructure, we now have the opportunity in front of us to become one of the largest data center companies in North America via the conversion that was just discussed of up to 500 MW of our existing infrastructure to high-performance computing.
With the contracts that we signed today, total revenue, cumulative revenue over the life of the contract represents $6.7 billion, which will put us in a very strong position from a cash flow perspective, to be able to continue to grow that side of our business. So we'll jump right in. Forward-looking statements, read them at will. So right now, we're in a very strong position. Thank you. We're in a very strong position. We've signed 382 MW of high-performance computing contracts, and we are in one of the leading market positions in terms of Bitcoin mining, in terms of bitcoins mined every single day. And so that put us in a very strong position in Q2, where we were very effectively managing a post-halving world. The post-halving world is very challenging, and what we're seeing today is even more challenging.
Average hash price in Q2 was about $0.062. The average hash price that we're seeing over the course of Q3 so far is somewhere in the fours, so we're in a much different paradigm shift in terms of what it means to actually be an efficient Bitcoin mining company. Right now, we're currently expanding our Pecos, Texas, our West Texas site for Bitcoin mining. We mentioned we will complete another 100 MW over the course of this year, so a great way to think about the business today is that we have 1,200 MW, of which about 800 has already been completed and is fully operational. And we had about 600 MW dedicated to self-mining, another about 200 MW dedicated to hosting.
We're sunsetting our hosting business over the rest of this year, and we're actually pulling back from 600 down to 500, but at the same time, we're refreshing a portion of our fleet. That'll give us the opportunity to continue to expand our hash rate, and then we just completed in August a convertible note raise of $460 million. That provides the opportunity to eliminate a significant amount of high covenant, high collateral debt that we had on our balance sheet, so we eliminated about $275 million in debt on our balance sheet, so we were able to add about $175 million of cash to our balance sheet, alongside of essentially restructuring our entire capital structure, which frees us up to go out and do more interesting deals going forward.
This lays out some of the things that we've been up to this year. What's not on this page is we emerged from Chapter 11 in January of this year. It's only been about eight months since we've emerged. Shortly after emergence, we signed a 16-MW deal with CoreWeave for our in our Austin, Texas, data center. That was for 16 MW. Shortly thereafter, we began discussions on a much broader conversion with CoreWeave and a number of other parties. Since that time, we signed a 200-MW contract June third. We signed another 70 MW a few weeks after. We signed another 112 MW shortly after that. Putting us at 382 total MW today, that puts us up there with many of the largest data center companies out there.
We've also announced our Pecos development, which is the 100 MW, adding on to the 71 MW we already have operational at that site, and we've been able to recapitalize our entire balance sheet, so we came out with a very complex capital structure, included a number of different warrants, included multiple forms of take-back debt and new notes to many of the previous creditors, but we're in a position today where we've been able to eliminate all of that previous debt that we have left over from the Chapter 11 process, and we're able to put ourselves in a very strong cash position alongside of all of our warrants now being available to exercise, so this is a good highlight of our total footprint today, so we're operating currently about 830 MW.
We have 1.2 GW contracted and built out high-voltage infrastructure across all of our sites. And so we currently operate in five states, with a sixth state today under development. That's Muskogee, Oklahoma. And so we have three sites in Texas, we're in North Carolina, we're in Georgia, we're in Kentucky, we're in North Dakota, and we're in Oklahoma. And so we have a very broad footprint, and that's actually providing us a very interesting opportunity in the high-performance computing side, where we have low single-digit latency to many different major metropolitan markets, which is incredibly important when you're trying to sell high-performance computing. Because one large tech company might have a need for 50 MW to service a very specific geographic area.
Having our broad footprint, and as we're looking to add more sites to our footprint, we're looking at more of a geographic spread to be able to service more types of clients and more client demands. So this is, I think some people have been calling this the racetrack page. Since 2021, we've mined more Bitcoin than any other publicly traded mining company. This really comes down to the fact that we were the first institutional Bitcoin miner in North America. We had hundreds of MW already developed prior to many of the companies that exist in our industry today even being founded.
This really highlights that fact that we've been operating at scale, dating back from, you know, two halvings ago, which is something that's very unique in our industry, and something that many other companies have really only gone through the previous halving that we just went through. This is just another touch on 2024. You know, we signed the 382 MW. We talk a lot about HPC, but something that we haven't talked about yet was the signing of our contract with Block. The Block deal is very interesting. You know, you look historically at this industry, 20%-25% of all of your revenue goes to Maintenance CapEx. That creates a very challenged business model on a long-term basis, unless your cost of capital is near zero.
And so the way we're approaching this is, if we expect form factor to change, essentially, that shoebox that we've all been building or we've all been hosting for the last decade, is that the optimal form factor for delivering long-term returns in Bitcoin mining? We don't believe it is. And so the deal that we signed with Block is part of a broader thematic shift in Bitcoin mining, where we believe that machines will start to look much larger. You'll be replacing only hash boards in the future versus entire machines, so chassis, power supplies, and all the other necessary components. And we'll also be able to operate much more efficiently inside of the facility. So cooling is going to become much easier in a new design.
All of these different components added together are going to add up to lower CapEx on an ongoing basis, faster refresh cycles, and more efficient data centers. Our goal is to bring what has historically been 20-25% of revenue to maintenance CapEx, bring that down into the mid-teens, to present, to really give us an opportunity to be competitive long-term in this industry. Because right now, you know, we're looking at a hash price at $0.04. Extraordinarily challenged economics on a long-term basis. ROIs and machines are continuously expanding, and if there's any bump up in Bitcoin price that may drive a higher hash price, we believe that the opportunity to capture that, based on having a more flexible footprint of machines, is actually going to be one of the biggest contributors to profitability on an ongoing basis.
Having the ability to capture those three to four months a year, where there's above-average economics, is incredibly important. That's something that we're really focused on getting to. Just to touch quickly on Q2 financials. We produced $141 million in revenue. We did have a $805 million net loss that was entirely driven by mark-to-market on our warrants. So something unavoidable on our balance sheet, based on the fact that our stock price increased, and so the liability or the non-cash GAAP mark-to-market is purely related to the increase in those warrant prices, and is a fully non-cash item. But we also did produce $46 million of Adjusted EBITDA, which is something we're really proud of in the first post-halving quarter.
So this is our revenue mix from Q2, and it's something that is going to change. So I'm going to touch on two things on this page. The first is the percentage of our business is self-mining. Now, by the time we get to the end of twenty twenty-five, we do expect this revenue mix to change pretty radically. Over time, we're going to continue to be ramping up HPC MW, which is going to continue to push as a greater percentage of our revenue mix. The second point on here is the HPC gross margin.
You know, that was the Q1 that an existing facility, and it's actually one that we're leasing, it's the only facility in our portfolio, we're leasing an existing data center to re-lease it out based on an immediate need that CoreWeave had, but that's driven by a mismatch in revenue recognition versus cost recognition due to GAAP. And so we, we've said in our 2Q call that we expect 35%-40% ongoing gross margins from that business beginning in 2025, when that GAAP revenue cost mismatch begins to unwind. So this is our, our self-mining fleet composition as of today. This is something that will shift as we begin to shrink our footprint for self-mining.
It's also something that's going to shift as we begin to refresh small portions of the fleet prior to the introduction of the Block chips in 2025. Today, we sit in a very strong position. We're at about 24 joules per terahash, and as we look to 2025, this is really where the acquisition of the Block chips are really going to come to fruition, where we're able to continue to expand our hash rate, while also bringing down our joules per terahash significantly. This is ... I almost want to title this page Not Complete, because this doesn't include what we just did in August. We started the year, actually, January 1, we started with $1.1 billion in debt prior to our emergence from Chapter 11.
We eliminated just over $400 million of that to get down to the $608. So January twenty-fourth of this year, we're sitting at $608 million in debt. Over the course of the first two quarters, we were able to bring that down to about $550. Now, shortly after the end of Q2, we hit the mandatory conversion of the convertible notes, which was a stock price threshold. That eliminated $260 million in debt from our balance sheet. And then, that provided us the opportunity to go out and do a refinancing of the rest of the debt that was on our balance sheet. We did have some holdover of some of the existing debt that's very low interest rate that we decided to keep on balance sheet.
But for the takeback or for the senior notes, $150 million delayed draw term loan, $61 million, so that's $211 million. That was given to the previous convertible note holders. We're able to refinance all of that, alongside of also refinancing the minor equipment debt. So we're essentially able to eliminate all the collateralized debt on our balance sheet, while also removing all of the covenants, which put us in a very strong position, alongside of putting about $175 million in cash on the balance sheet. Which is gonna allow us to continue to expand our high-performance computing footprint by acquiring new sites, and then also refreshing and paying for newer machines to continue to lower our joules per terahash and increase our hash rate. So this is, this is a slide that we're pretty proud of.
You know, coming out of a post-halving world, you expect cash cost to mine to increase by 100% if difficulty stays the same. Now, in Q2, we actually came in at about $29,000. So in Q1, we were at, I believe it was $18,900 total cash cost to mine. And so you'd have expected us to come in in the high 30s in terms of a cash cost to mine in Q2, where a majority of the quarter was actually in the post-halving world. So we only had 3 weeks pre-halving.
We came in at $29,900, which is something that is very impressive, and it comes down to our operational capabilities and the playbook that we laid out prior to the halving by saying, "You know, this is how we're gonna begin to, allocate our power. This is the power programs that we're gonna participate in post-halving. This is how we're gonna move machines throughout our facility footprint to lower our break-even hash cost." All of those components, including the fact that we run our own firmware, our own energy management software, and our own fleet management software, all of these components added together resulted in a much lower cash cost to mine than what our peers experienced on a quarter-over-quarter basis. All right, so now the fun stuff.
So, this is the illustrative economics of our HPC contracts that we've signed with CoreWeave so far. So it's structured in a completely different way than how the data center industry has structured contracts in the past. So if you take a traditional data center company, you're looking at 70%-80% debt financing or 70%-80% loan to cost. You're looking at 20%-30% equity checks to pay for the remainder of the project. Essentially, what we did is a bit flip it on its head. The debt portion of it is being fully funded by our customer. That does not have to be paid back. And the equity portion of that check is what is prepaid by CoreWeave, and then we pay back via a credit.
And so when you look at this page, you see year one, year two. You see this CapEx credit. That is where our payback for the $1.5 million per MW that we have to pay back to CoreWeave flows through. So on a GAAP basis, we're gonna be showing the entirety of the revenue, but then when you actually work through the income statement, there's gonna be a credit back to CoreWeave, essentially a reduction in their revenue. So for the first two, about two and a half years, it's a little less than that, but for on average, about the first two and a half years, they're paying about 50% lower. It's no greater than 50% of total revenue that could be credited in any single month.
And so after we work through really the first 3 years of the contract, we expect long-term margins in the 75%-80% range, which is gonna be generating a very significant amount of cash flow. So it's on average about $1.45 million per year per MW, and we're expecting about 75%-80% gross margin. It's a cash gross margin, so it does not include depreciation and amortization, nor stock-based comp. But on a cash basis, expecting the cash flow over $1 million a MW on a long-term basis, it's a 12-year contract. So for about 9 years, we expect to do that. And then this is the delivery time schedule for the first 382 MW.
So what we're looking at here are 200 MW being delivered in the H1 of 2025, and then 70 in the H1 of 2026. Or sorry, 70 in the H2 of 2025, and then 112 in the H1 of 2026. Now, there are ramp-up schedules. This doesn't all turn on at once. So for that first data center, it's not 80 MW day one; it turns on. There's a ramp schedule between 10-20 MW a month, depending on the site, and so there is a ramp into full capacity. Now, we've announced so far that we have five sites that are currently under contract, so five sites are within this 382 MW.
However, we've not announced which sites, and that's mainly driven by the fact that right now, we're looking at additional power allocations across each of these sites. And so it will provide us the opportunity to grow our MW just within our existing footprint, outside of the sites that we're currently going after. And on an ongoing basis. I believe it's the next idea. So on an ongoing basis, right now, I'll touch on the first point first. You know, we have a hundred and eighteen MW of the five hundred total that we've allocated, currently under option agreement with CoreWeave. We just announced that we extended that option another thirty days.
You know, there were points on the 32 that had to be ironed out, given the tight delivery schedules on certain parts of those MW, that we had to confirm between us and CoreWeave to ensure that we can meet those timelines. And so now we've refocused those discussions back on the 118. But really, these second two points are the future of the company here, which is evaluating or executing on the existing pipeline of sites. And so we've been evaluating GW of sites just in the United States over the course of the past six months. This is something where we look to get sites under either option agreement or exclusivity under LOI, in order to give us time to then go sell that to an end customer.
So it's not. We're not looking at being speculative on sites. But the third point on this page is diversifying our customer base, which we're essentially sold out on the first 500 MW of power capacity that we have available. And so as we look forward to new site developments, we're looking aggressively at bringing new customers into the business, and right now, the demand is insatiable for this power. And so it's provided us the opportunity to be much more specific about which sites will fit with each client. And something that we're very interested in right now is finding and getting a toehold in some of the largest tech companies, via potentially some smaller contracts, that have the capability to expand from maybe 25 or 50 up to 100 or potentially 200 MW.
We think that's a really interesting opportunity for us to continue to expand our footprint. And we've laid out some of our 2024 goals. You know, we're expected to get to 21.8 exahash before the end of the year, and that's with the migrations of five of our facilities to high-performance computing. We have 16 MW already operational today on the high-performance computing side. We expect to have about 800 MW fully operational. Now, there's some give and take on that number. And the reason why it has a tilde is because we're turning off big sites, and we're moving machines to other sites. That's why we've rapidly increased the timeline on our Pecos location, so to bring on another 100 MW.
And then, one of the great parts about some of these conversions is that some of our best locations for high-performance computing were also some of our highest power cost sites, which in the future, power for high-performance computing is going to be full power pass-through to our clients, but it's actually lowered our average power price on our existing Bitcoin mining sites. So it's very good tailwinds going into a more challenged hash price environment, which providing us the opportunity to actually pull in our power cost, which is something that you saw reflected in our cash cost to mine at about $25,000, purely on a power basis. So just to close it out before we open up the floor to Q&A, you know, we're in an incredible position today.
We've allocated and already signed contracts nearing 400 MW, which will put us as one of the largest data center companies in North America. We've also been the largest Bitcoin mining company dating back to 2021, a business that we execute on very, very well. We've solved our capital structure issue and put us in a very strong position for continued growth on both sides and diversifying our client base, which we believe puts us in a very strong position for continued shareholder value creation over the many coming years, so with that-