Good morning. Thanks, everyone, for dialing in. It's May 16th. This is our 3rd day of the Needham Tech, Media, and Consumer Conference. Very excited to have everyone joined us so far in these first couple days, and excited today to have Core Scientific presenting. We have their CEO, Adam Sullivan, and CFO, Denise Sterling, on for us. They are going to present for, I think this is in total a 40-minute session, so about 30 minutes-35 minutes, and then we will take some Q&A from the audience at the end for five or so minutes. With that, Adam, would love to let you dig into it.
Sounds great. Thank you very much, Joh n, and thank you for including us in our conference. I just wanted to start off by saying, thanks to everyone that's joining us today. And I think for us, you know, we want to start at the beginning of our journey. First, we'll start with our safe harbor statement before we kick in. So flipping to slide three here, when Core Scientific began, our founders recognized the importance of infrastructure and that access to steady, high-power sites would become more valuable over time. They started the company as a hosting business, hosting actually other Bitcoin miners, and realized that operating at scale offered very important economic advantages.
So if you look back at the history, we started by hosting Bitcoin, other Bitcoin miners, then we began hosting GPUs at two of our data centers and actually built a tier 3 data center inside one of our facilities. And then, as our footprint began to grow, growing from 100 to 200 to 300 MW, you know, we are on a growth path to actually take this to a much, much larger scale. So in 2021, we started to shift to more to self-mining, actually. So from hosting Bitcoin miners to self-mining and continued expanding the footprint. We actually were the first also to achieve 500 and then 700 MW of operational infrastructure. Today, we operate 745 MW of infrastructure at seven different sites, currently split about 77% self-mining and 23% hosted machines.
We've also began, and we announced this previously, but we began and paid for the next 72 MW of operational or of infrastructure at our Denton, Texas, site. That's included in the total of the 372 MW of partially developed infrastructure that we have. Upon full completion, that would actually take us to about 1.1 GW. We produced more Bitcoin than any other public mining company since 2021, and we're one of the most productive Bitcoin miners, both from a hash rate utilization perspective, essentially, how many bitcoins do you get out of each exahash of operational machines, but also on a joules per terahash basis, recently bringing our average joules per terahash below 26 as of April thirtieth.
And as we speak today, we've actually announced that we're in the process of going through negotiations on about 5 MW of our infrastructure out of that total of 1.2 to conversion to HPC facilities. So all along, this company has been transforming energy into high-value compute with superior efficiency at scale, and Bitcoin mining really led us to this valuable position. It helped us build a platform of infrastructure assets. It helped us build one of the strongest teams in the industry, mainly driven by traditional data center people, and it is really the platform upon which we will continue to grow this company. So let's go through some of the market-leading stats for Core Scientific. We've been the top Bitcoin producer since 2021.
We've been delivering strong financial results over the past few quarters, including Q1, where we actually produced a very significant amount of adjusted EBITDA, posting $88 million. We're the largest infrastructure owner by megawatts. We currently operate about 20.4 exahash for ourself, and we're currently growing our infrastructure footprint. So if you look at where we are today, by the end of Q2, we should surpass 800 MW of operational infrastructure. We're also diversifying our revenue base. We previously announced a transaction or a deal with CoreWeave on 16 MW to host GPUs on their behalf, and we're generating cash flow out of our operations, and we're going to continue to grow our operations out of our operational free cash flow.
Kind of the last point here, which I think is one of the most important, given the explosive growth in the data center industry today, we already have our team to grow to really the next step and grow into HPC. Talent right now, given the growth trajectory of that industry, is extremely competitive, right? If Morgan Stanley, Goldman Sachs, all the bulge bracket banks are saying that data center growth is going to double over the course of the next six years, that means there's a lot of folks that are going to be very well employed over that course of that time period.
And luckily, we've been able to bring together a very talented team to actually help us become one of the largest Bitcoin miners in North America and will help us continue to grow into the HPC vertical. All right, so let's talk a little bit about our sites. So today, this is what our 745-MW footprint looks like. You know, in Q1, we averaged $0.043 power price across our portfolio. We're currently operating at least 100 MW at all of our sites. And so these sites are actually spread across five different states. And so when you, when you look at this map, we do currently have seven facilities. We have two in Texas, and we have two in Georgia.
And so some of our data centers, when we talk about this growth of HPC, some of our data centers will continue to support Bitcoin mining. However, some of them are actually very well positioned for alternative compute. And the main component here that's really important when people come and talk to us about potentially converting facilities to HPC, you know, there's a number of major factors here. So there's the fiber lines, low latency to major metropolitan markets. There's the current electrical infrastructure. The great part is, we have insight into all of the different areas that are necessary in order to build out a more traditional data center at each of our facilities. And actually, one of the main components here is that we have insight or already have the fiber lines at many of the facilities that we've identified for HPC.
There's one site that's not listed on here. It's our Oklahoma site. That site actually has 100 MW of power, and that's not included in the 745 MW. And so on top of the 745 MW, we have 372 MW contracted power in Texas and then another 100 MW of power contracted in Oklahoma. So this page, we think, really illustrates our lead in this industry and where we've been over the course of the past two years. You know, it's not just been on the infrastructure side, but also on the Bitcoin mining side. You know, we've earned nearly 37,000 Bitcoin since 2021, and over that time period, we've installed well over 700,000 machines in our operations from nearly every manufacturer.
I believe the number is about 16 manufacturers that we've had operating in our facilities in the past. So we've been able to inform all of our software stack. You know, we're the only company that's built all three levels of our software stack, so energy management, fleet management, and our own firmware, but it's also informed by our design. We're on Gen 5 of our existing design, and that comes down to including all of these different instruments in our facilities to help us make smarter data centers. So we've built our own sensors that help us monitor pressure, humidity, weather, wind. All of these variables actually come deeply into play every single day in our facilities, and we're actually able to tune our machines according to prevailing winds, or humidity, or pressure in a facility.
That's something that is incremental gains that we're able to get on top of our competitors. So we're able to eke out additional percentages of hash rate out of the same machine that our competitors are running at spec rates. We're actually able to gain more terahash without necessarily even pushing up the efficiency, and so that's something that we're really proud of, and it's something that does give us a competitive edge in this industry. And so this is one of the kind of the proofs in the pudding here. You know, when you compare us to our scaled peers, so think Riot, Marathon, CleanSpark. On a per exahash basis, we're mining about 9 more Bitcoin per energized exahash than they are every single month.
When you compare it to our non-scaled peers, so folks running smaller data centers, you know, it's much easier to operate a much smaller data center when you only have, you know, a few data centers. And so we're still mining nearly 4.5 more Bitcoin than our non-scaled peers. And so this really illustrates our competitive advantage. The fact that we're able to run with higher up times, that we're able to actually refine all of our firmware according to prevailing weather patterns, et c. All of those things are incredibly important for us as we look to continue to scale our mining business. And so this is something that not only is noticed in the Bitcoin mining industry, but this is actually also something that folks in the traditional data center industry have noticed as well.
You know, it goes back to the fact that we've hosted companies like CoreWeave in our facilities, hosting GPUs dating back to 2019. And so our learnings, not only on the Bitcoin mining side, but also on the GPU side, this rigorous process that we go through in development, all of that has led us to being one of the best operators in this industry. So talking about Q1, we delivered really strong results in the first quarter. We led all Bitcoin publicly traded Bitcoin mining companies in the first quarter. We improved our miner efficiency to 26.85 joules per terahash. You know, that's something that when you look to April, we actually got below 26, and we produced $179 million in top-line revenue. That was at a 42% gross margin.
$150 million of that was from, from self-mining. We generated $211 million in net income. That did include certain one-time items, but we did produce $55 million in operating profit. I think one of the best parts about Q1 is we were able to put significant cash on balance sheet out of our operations. So we were able to put 98, or we were able to end the quarter with $98 million cash and equivalents. And at the end of, right after the end of the quarter, we actually paid down $19 million in debt, and that brought us below $600 million in debt, which is something that, you know, I think is a, is a good hallmark for us, continue the opportunity to reduce debt where possible.
And that's something where we had to pay the $19 million, actually, to continue construction at our Denton facility. So those related to mechanic's liens, and the project actually only cost us $1 million to complete. And so it really opens up the Denton facility to continue construction past that 72 MW. We also achieved a very favorable cost to mine, so $18,915 per Bitcoin. And that's something that in a post-halving world, given some difficulty adjustments that we've seen, you know, our expectation is that, you know, it would be somewhere in the mid-$30,000 range based on where, where things are today. We also delivered a 16-MW HPC data center to our client, CoreWeave, which was more than 30 days ahead of schedule. So this is really where Core Scientific is going right now.
You know, we have a huge opportunity ahead of us to actually capitalize on all the own infrastructure that we have for HPC. You know, the market drivers are evident, you know, 160% expected growth in U.S. data centers from 2023 to 2030. But really the main point is here, when people ask us what the competitive advantage is, it really comes in two areas. It comes from the infrastructure and the platform that we've built. You know, having 1.2 GW of contracted power with a high voltage there, with 745 MW of operational facilities, those are all incredible opportunities for traditional HPC companies to actually work alongside of us and partner with us to build out facilities.
The other part is, though, I go back to the fact that this industry is experiencing such rapid growth right now, the HPC industry. And if data centers are truly going to grow over 100% over the course of the next six years, that requires a massive amount of talent. And that's one of the main reasons why people are coming to Core Scientific, is that we have the talent already in-house. And so that's what excites people. They see we have the infrastructure assets, and we have a known team in the data center industry that people want to work with. And so what that means for Core Scientific is an incredible opportunity over the course of the next three to five years to build out a very competitive business.
If we build out all 500 MW, it would put us in the top 10 data center companies in the United States. What does that mean for our shareholders? Well, multiples in this, in that industry are much higher, especially if you're in long-term, stable offtake arrangements with, with clients. So what we're talking about there is, you know, let's call it 10 years or greater with renewal, with renewable options. That's something where, you know, we're targeting today in terms of how we're thinking about this business. It brings predictable, stable, recurring cash flows to the company. It diversifies our customer base and our revenue mix. From our perspective, at least how we're thinking about the business today, it provides us the opportunity to continue to expand our Bitcoin mining business as well.
Because we're going to have a stable cash flowing business in our company that's going to allow us to actually make countercyclical investments in Bitcoin mining as well. And so when we think about this business as a whole, it's really one of the most exciting businesses and opportunities that we've seen come across for having high-power sites alongside of major fiber lines. And so that's really one of the most exciting things that we're seeing today, and we're going to continue exploring and negotiating with potential clients over the course of time in order to find the right clients, the right groups, and we'll talk a little bit more about the structure in a moment in terms of what we're targeting. All right, so talking about Q1. So we ended the quarter.
You know, I talked about the 77%, self-mining as a percentage of miners. On a revenue perspective, we produced 84%, 84% of our revenue came from self-mining, 16% came from hosting. And what you can see is our self-mining business over the course of the past year, it's actually increased from a gross margin perspective. So we've gone from 26%- 46%, and we've actually been able to increase the gross margin on our hosting business as well. You know, this comes down to the fact that we've actually been able to hone in our hosting contracts and make a very profitable business out of hosting. So we've targeted new segments in the market. We were one of the first companies to do profit sharing or proceeds sharing arrangements with our clients.
That helps us and our clients, because in down scenarios, yes, it contracts our proceeds sharing arrangement, but it also keeps our clients profitable. And so we don't have to worry about refilling those slots during time periods where Bitcoin may be pulling back or economics may be challenged. But during up periods, it also provides us the opportunity to actually participate in the upside of Bitcoin. And so that's something that we're very excited about as well. You know, growing our mining revenue by more than 50% year-over-year, and our hosting revenue by more than 25% year-over-year, is something that, you know, we're going to continue to focus on. You know, continuing to be able to grow the revenue segment of our Bitcoin mining business.
So what comes into play here on the self-mining side, you look at our fleet composition. You know, as I mentioned earlier, today, we're under 26 joules per terahash. You know, this comes down to. This is our fleet mix as of quarter end, so not as of April 30, but as of March 31. The opportunity here is continuing to refresh into either S21s or XPs. You know, what we're seeing in the market today is market pricing for anything that's not an S21 has dropped significantly. However, the joules per terahash between an S21 and XP are not significantly different from the perspective of what are the returns on those machines.
There's opportunities that exist in the market, and this is just given the broader displacement in the market, where a number of companies are operating at very high costs sites. And so they have to upgrade into the S21 era to maintain profitability, to continue producing Bitcoin. And so that actually creates opportunities for Core, where we have cash on balance sheet and the ability to deploy it and actually find machines that we can install now. You know, another part of this is, you know, you look at our machine fleet mix, and a lot of people ask: "What is the opportunity inside of your fleet if you were to refresh all of your machines?" Well, if we refreshed our S19 Pros and our traditional S19s or our base S19 models, that would increase our exahash by more than 10 exahash.
So we would increase our operational exahash by more than 50% if we were to do that. The other part is we've been working with a number of different mining manufacturers. Not all existing mining manufacturers, but new mining manufacturers that are emerging. You know, we think 2025 is really gonna be the year of democratized hash rate, where more mining manufacturers are coming to market. And we're one of the most experienced miners out there. You know, operating more than 700,000 machines in our facilities gives us a lot of learnings, not only on the operation side, but also on the repair side.
We're looking at some of the newest ASICs that are gonna come to market over the course of the next year, and creating our own solutions for those types of ASICs that actually provide us the ability to generate greater productivity from our Bitcoin mining infrastructure. All right, so we thought this would be helpful for investors. You know, it was part of what was a little bit confusing for investors was that our Q4 filing, we were still in our restructuring process. When you look at 12/31/2023, or the end of 2023 to Q1, this was our reduction in debt. So we went from about $1 billion in debt to about $608 million.
Now, given the $19 million pay down, that actually comes below $600 million. Really, most of this reduction in debt was from equitizing notes and loans. The main part here is our entire capital structure was actually built to have optionality and time. You know, those are the two components that are necessary, or if you're gonna have debt on your balance sheet as a Bitcoin miner, given the volatility. We've built optionality via a tranche one warrant, which, upon full exercise, brings $670 million in cash into the business. Then the other part is we also have a mandatory convertible note. That mandatory convertible note has a mandatory trigger at a $779 price, and so that reduces debt by $260 million.
And so reducing all of this debt actually provides us the opportunity to be in a positive net cash position of about $400 million. And so right now, we're extraordinarily focused on growing our share price from the perspective of the share price actually does dictate really how much debt we have on balance sheet going forward. And so we believe, based on continued execution of our business, you know, consistently putting up quarters like, what you saw in Q1, is a great opportunity for us to continue to show the market that we execute, and we execute at the highest level. So flipping to the next page here. You know, this is our direct cash cost to mine, and this mainly consists of power in our mining facilities and any of the associated costs with our mining facilities.
So there's really two ways to look at this: There's your cash cost per Bitcoin, and then your cash-based hash cost. From a cash cost per Bitcoin in the first quarter, we came in about $18,900. This positioned us as one of the lower-cost miners in the space. The other side of this is the cash-based hash cost, so coming in at $0.0326. This is something that does not double in a post-halving world, and that's something, you know, holding all the variables constant, and that's something that I just wanted to take a moment to explain. So when you go through a halving process, you're mining less Bitcoin. Ergo, your $18,900 actually increases.
Your cash-based hash cost, though, is spread over your terahash pool, and so the cash-based hash cost actually stays flat in terms of when you think about hash price. Hash price did come down in the post-halving world, but our cash-based hash cost stayed the same. So let's talk about 2024 and what our kind of goals and targets are here. So our targets are 21.8 exahash by year end. You know, today, we're sitting at 20.4, so that will require some new machine acquisitions, whether that be for filling out new capacity or refreshing some of our existing machines. We have the 796 MW of owned operational infrastructure. You know, to provide some color around that, you know, we're currently operating 745 MW.
We have plans to add another 72 MW during the course of Q2 of this year. Including that, 745 MW that's operational today, our 21 MW actually of what we call kind of our flex space. So it's where we take older generation units and put them on separate meters from the rest of from the rest of the operations, and we actually run them on different power strategies. So the reason why we do this is, let's say we can operate a machine at 70% or 80% uptime but generate more profitability on older generation units than what we could sell them for today, you know, we're always gonna take that opportunity to maximize the profitability of those units.
And so by year end, we don't know where profitability will be on 34 joules per terahash machines, and so that's why we're at the $7.96. And then we're saying that over the course of this year, we'll be at an average power price of about 4.5 cents-4.7 cents per kWh. So just to summarize here, we've been able to achieve remarkable operational consistency. It's something that people noted even over the course of 2023, and when you look back through all of our reporting back to 2021. We have remarkable consistency, delivering consistent Bitcoin production with our fleet.
We've been one of the largest mining companies and one of the most productive public Bitcoin mining companies since 2021, and that's something that everyone in this industry understands how strong of a team we have, and we have one of the strongest infrastructure platforms in the business. Now, that infrastructure platform positions us very well for growth into HPC by addressing some major imbalances that we see today between supply and demand of available infrastructure, and especially with high-power sites. You know, sites over 100 MW are very rare to find these days, because it's not just Bitcoin miners anymore competing for those sites. It is the likes of all of the major tech companies and data centers.
From our perspective, you know, we believe the market does not appreciate the value proposition, you know, based on when you look at the valuation multiples across our industry. We believe by executing our plans, continue to raise awareness around what our competitive advantages are and what the market opportunity is for us to execute on, we believe we'll be able to close that valuation gap, which will be a major benefit to our shareholders. I thank you for your time today. John, I'm happy to take some questions.
Yes. Sorry, I was trying to find my, the mute button. Great. Thank you, Adam, appreciate that. That was, I think that that was a great overview, and some of the stuff we got a little bit more details. Love the slide about, your, your different rigs, and the, joules per terahash for each. Great! So, I will look at some of the questions here. I do have one or two of my own, and then we'll, we'll get a couple, from the audience going. But so from my end, first one on HPC. So obviously, big ambitions there. I guess just first off, 'cause we do have a little bit of time, talk to us about kind of how the overall strategy has shifted a bit, where you guys want to go after that segment.
A lot of Bitcoin miners are doing it. You guys have done it more aggressively than the other ones, I would say so far. So, talk to us a little bit about that, and then I have more detailed follow-ups on that.
Yeah, so I think it's important actually to bifurcate what our strategy is in this industry versus other, Bitcoin miners who are getting in. So when you look at some of the competitors, they've gone it from the approach of a, of acquiring the GPUs and operating them themselves. So, you know, the largest installation, I believe, of GPUs amongst any public Bitcoin miner is about 1.5 MW. Obviously very, very small, and they're running. You know, the 1.5 MW deployment, I believe, is all GPUs. You know, there's another one that's about 1 MW.
You know, these are companies that are selling short-dated contracts for compute to non-investment-grade off-takers, which will be very hard to find the valuation multiple uplift from short-dated contracts when the ROI period, they're being fully employed, the entire time period is at least two years. And so when you think about the investment on a per megawatt basis, new GPUs will cost you anywhere from $25 million-$35 million per MW, and so obviously that business is extraordinarily competitive. There's very high barriers to entry. Just getting the queue on cards is very difficult. So the way we've approached this, and you know, you saw it with our initial deal with CoreWeave, is we have exceptional talent, we have exceptional expertise in this industry, and we have infrastructure assets that we've been in discussions with potential clients on for a while now.
The way we're approaching this is we're actually going to develop this. We've been working through designing, developing, and determining what are the capabilities of each of our facilities. We've been working with our potential clients around what the designs look like for all of the facilities and what their requirements are. When you think about the next generation of GPUs that come out in 2025, 2026, you know, those have different requirements than what traditional data centers can provide, and so this is something where they need to find more application-specific digital infrastructure to actually run these at the most efficient level.
From our perspective, if we have the talent and we have the infrastructure assets to execute on this, and we have clients that are willing to put up the CapEx to actually refit these facilities, then this is an incredible opportunity for not only the company but for our shareholders, where we can deliver predictable, stable, recurring cash flows out at minimum, let's call it 10 years. That's something that we're taking this purely from a digital infrastructure approach, not from the approach of trying to go out and acquire GPUs for ourselves.
Got it. Okay. But so, so talk to us a little bit about that CapEx angle you mentioned, 'cause.
Mm-hmm
We cover other companies who have really leaned into this, and building out those data centers are a massive CapEx lift. A lot of financing needs to come together for it, and that seems to be one of the main bottlenecks. So talk to us a bit how you are thinking about financing, and I mean, so for a 100 MW site, we hear that should be, like, $800 million. Is that fair? Kind of $7 million-$8 million in CapEx for a 100 MW HPC site, and then talk to us about financing.
Yeah, of course. So on our Q1 earnings call, we actually talked through some of the CapEx numbers. So we've been able to dial these in based on all of the design, working through supply chains to determine the timelines for when things could be delivered. So based on our current footprint, based on the current assets we have on site, it would cost us between $5 million-$8 million to convert per MW. This is something where we've said to all of our clients, we took a much, much larger list of potential clients, and we said to each of them, "You know, you have to be able to pay for the CapEx." And that's something where that narrowed the list down to just over a handful of potential clients, and the response was absolutely.
From a financing perspective, we view this as a very great way for actually to get a very low cost of capital financing structure. Partnering with our clients on all the design choices, on all of the, all of the timelines, and also partnering with them on the capital side, also forces all of the designs to be much more capital efficient from the perspective of, you know, we're gonna build the most efficient operations possible, but also build it to actually fit what's coming in the future in terms of GPUs versus what's here today. Because what's coming in the future is gonna be much different in terms of structure, in terms of infrastructure needed, in terms of the cooling needs.
Those are all gonna be different, and so that's why we work very closely with both potential clients as well as major GPU manufacturers to actually design the right thing, not just what's in the market today.
Got it. Understood. Thanks for that. I have one more, and then we'll get to a couple audience questions that have come in. You know, so besides financing, you do hear about supply side issues in not necessarily on getting the GPUs themselves. It seems that folks have been able to get that, and I know that's not even necessarily where you're targeting. But what are some of the supply side constraints on that side of the business? And do you see the, or have those been lessening, increasing? Any thoughts there?
Hey, John, I just want to make sure I understand. So you're saying, what are the constraints on the infrastructure side?
Yes.
Yeah, so this is something where we've done a very deep dive into the supply chains related to this, the build-out of this business. You know, we put out on our Q1 call, we said it would take 3 years-4 years to fully complete the 500 MW. Now, that takes into account supply chain constraints that currently exist in the market. You know, some of the supply chain constraints that we're seeing, you know, we already have all of our substations. You know, those right now, if you go out and try to buy them, it'll take you at least probably 36 months in order to land those.
But what we're seeing, the major supply chain constraints is really on the generator side, a bit on the dry cooler side, and a little bit on the battery side, so the batteries that are specific, specifically built for data centers. And so those are kind of the main components that we view as long lead items in this. The nice part is, you know, as we continue to work with each of the potential clients, the interesting part is many of them are actually already sitting in the queue on some of these longer lead items with a recognition that they knew that they were gonna have to acquire those items in order to build new capacity.
And so being able to not only tap into the supply chains that we've been working through on the Bitcoin mining side for the past seven years, but also having the opportunity to tap into these larger tech companies that have much larger supply chain footprints and the ability to actually get in the queue and have been in the queue for quite some time, you know, that has an opportunity to actually pull forward infrastructure development on a faster timeline than what we're even looking at today.
Got it. Okay, that's helpful. Thanks for that. Okay, cool. So let's. We have about seven minutes, so I'll turn it over to some audience Q&A. First one asks, "What could we expect for operating and EBITDA margins for HPC business?
Yeah, so that's, this is actually something we also talked about on our Q1 call. So we talked about a 75%-80% gross margin, and so, you know, that really almost directly relates, you know, minus some potential OpEx, really directly relates to what we would expect to see, you know, kind of flow into the bottom line. And so that's something that, you know, we view as a high margin business with stable free cash flow generation, and so, you know, we're, we're really excited about the opportunity.
Do you, you know, is that margin profile changing at all with supply coming online, or there's just so much demand that it's, it's still, you know, you still have kind of pricing power there?
There is still significant pricing power. What we're seeing in the market today is actually interesting. So companies that are doing more AI-related applications and have to build out, you know, the newest infrastructure for the GPUs that are coming out, whether it be next year or 2026, they have nowhere to put these machines, right? And from NVIDIA's production perspective, there's gigawatts that are coming to the U.S. of these cards, and so that's something where, you know, multiple gigawatts of data center infrastructure are gonna come in the next two years? There's not many sites that are being built out that didn't already have clients from the major data center companies. You know, data center companies are building these large facilities, but they're doing it with clients in hand, and so they're building out three years with clients in hand.
And so for them to actually go back and either cancel contracts with existing clients is not something that they do, right? They're valued so highly because they have very large investment-grade counterparties as off-takers that are signing very long-dated contracts. So even the existing facilities that they have, they don't have space in them. And so everything that we're seeing going forward is an opportunity for us to actually capitalize on tech companies that desperately need space over the course of the next three years, while they already have delivery schedules for GPUs but don't have a home to place them.
Yep. Got it. That, that's interesting. That's helpful color. Okay, the next one, this one asks, kind of back to Bitcoin mining. It asks, "So hash rate has come offline. Do you think early innings of that, or has hash rate bottomed here?
I don't necessarily think hash rate has bottomed. I think there's more challenges on the horizon. The reason why I say that is, at a $0.05 hash price, you know, if we drop down to even a $0.045 hash price, there's a number of operators out there today that are very marginal in terms of what type of cash flow they're actually producing off those machines. This is gonna be an extremely challenging environment three to six months from now, when those marginal operators begin to run out of capital. We also have what we see in the market is actually something really interesting.
We know for a fact that certain sites, even across the United States, from some of our competitors, are operating in a non-profitable manner, but they're going to continue to pay essentially more per Bitcoin than they're actually going to be producing, or what the spot price is of Bitcoin. So they're going to continue to do that because they need to show consistent Bitcoin mining production, and that's going to help them raise more capital via, you know, at-the-money programs, et cetera. If they just fall off a cliff in terms of what they're producing, that's a big challenge. I think what we're expecting to see is some of these marginal operators actually begin to take some of that infrastructure offline. Now, those machines will find new homes.
It's going to take time to move them because, especially when you think about the United States, the lowest cost operations in the world are operations where people are able to find extremely cheap power in very esoteric areas. And so they'll find. Some of those machines will find their ways to certain facilities across the world. It's going to take time, though, before S21s are deployed at scale, because many operations are still operating profitably, especially at the lower cost price point. So my expectation is that probably over the course of the next 2 months-4 months, we would have, you know, kind of a bottoming in terms of hash rate coming offline, and that's when we'll start to see a greater ramp in hash rate growth. Because what you're going to see is people who deployed capital into S21s actually have the ability to bring them online.
They'll go through their refresh on machines, and then by year-end, the expectation is, and depending on hash price, we'll see some of the older generation units actually start to kind of flip back online at new facilities.
Got it. Thank you for that. About one to two minutes left, so just this last one here. And this S in your pipeline, I believe it was around 1 GW, 1.1. Is there a, a sub-power cost level you're, you're kind of targeting? Like, is it some $0.04, or, or just kind of talk to us a little bit about the power dynamics of the, what's in the pipeline.
To grow from the 745 MW today to the 1.2 GW?
Yes, exactly.
Yeah. So 372 MW of that growth actually comes out of our Texas facilities. So between our Denton facility and our Cottonwood facility, which is located in Pecos. So that, those power economics are very attractive, right? The forward curve in Texas right now is higher than we would like. Currently, we're operating on index PPA at our Denton facility. That provides us the opportunity to participate in real-time markets, and what we're seeing are actually very attractive power rates there. You know, some of that's reflected in what you saw with the $0.043 power price in Q1 of this year, you know, actually coming in below what our expectations were for Q1. That also provides us the opportunity, when you think about a place like Cottonwood, where we can actually be much more selective.
We can do this at Denton as well, but we can actually run different power strategies that are meter-based. So we can run different efficiency machines on different meters to maximize profitability. That might mean that a 29 joule per terahash machine might be operating at, you know, let's call it between 90%-95% uptime to really choose a power price, and then at 34 joule per terahash machines, we could run them at a lower uptime to actually achieve a much lower power price. Because in Texas, it's really about avoiding the highest price time periods. And as you kind of work out those increments, you can actually lower your power cost by just avoiding some of the higher price time periods if you're running on an index PPA.
And so that index PPA provides us that type of flexibility to do so, and it allows us to actually elongate the life of all of our equipment, which is something that we think will provide much longer-term value to all of our shareholders.
Great. Thank you for that, Adam. We are at time, so I want to thank both of you, Adam, Denise, for joining us from Core Scientific.
Thanks so much for the time, John.
Of course.
Thank you, all.