Good morning, and welcome to Applied Digital's Third Fiscal Quarter, 2023 Conference Call. My name is Melissa, and I will be your operator today. Before the call, Applied Digital issued its financial results for the 3rd quarter of fiscal 2023, ended February 28, 2023 in a press release. A copy of which will be furnished in report on Form 8-K filed with the SEC and will be available on the investor relations section of the company's website. Joining us on today's call are Applied Digital's Chairman and CEO, Wes Cummins, and CFO, David Rench. Following their remarks, we will open the call for questions. Before we begin, Alex Kovtun from Gateway Group will make a brief introductory statement. Mr. Kovtun, please go ahead.
Thank you, operator. Good morning, everyone, and welcome to Applied Digital's Fiscal Third Quarter 2023 Conference Call. Before management begins their formal remarks, we would like to remind everyone that some statements we're making today may be considered forward-looking statements under securities laws and involve a number of risks and uncertainties. As a result, we caution you that there are a number of factors, many of which are beyond our control, which could cause actual results and events to differ materially from those described in the forward-looking statements. For more detailed risks, uncertainties, and assumptions relating to our forward-looking statements, please see the disclosures in our earnings release and public filings made with the Securities and Exchange Commission.
We disclaim any obligation or undertaking to update forward-looking statements to reflect circumstances or events that occur after the date the forward-looking statements are made, except as required by law. We will also discuss non-GAAP financial metrics and encourage you to read our disclosures and the reconciliation tables to applicable GAAP measures in our earnings release carefully as you consider these metrics. We refer you to our filings with the Securities and Exchange Commission for detailed disclosures and descriptions of our business, as well as uncertainties and other variable circumstances, including, but not limited to, risks and uncertainties identified under the caption Risk Factors in our annual report on Form 10-K. You may get Applied Digital's Securities and Exchange Commission filings for free by visiting the SEC website at www.sec.gov.
I would also like to remind everyone that this call is being recorded and will be made available for replay via link available in the investor relations section of Applied Digital's website. I will turn the call over to Applied Digital's Chairman and CEO, Wes Cummins. Wes.
Thanks, Alex. Good morning, everyone. Thank you for joining us for our fiscal 3rd quarter 2023 conference call. I want to start by thanking our employees for their ongoing hard work, including our construction and operations team for getting through the winter and keeping our facilities operational and construction timeline reasonable. Before turning the call over to our CFO David Rench for a detailed review of our financial results, I'd like to touch on some updates from our business over the last quarter and why we remain excited about the future and in our ability to deliver long-term, high-margin, sustainable cash flow. Let's start with an update on our two newest facilities, Ellendale and Garden City. We successfully completed energization of our 180 MW facility in Ellendale, North Dakota in early March.
This marks the second facility that we energize within North Dakota following the successful 100 megawatt facility in Jamestown that was energized in 2022. Recall that we broke ground on Ellendale in September 2022 and are now powering and operating the new site only six months after initial work began. This is a tremendous accomplishment given the harsh winter weather we experienced in North Dakota over the last several months, along with some construction delays. We continue to ramp up our capacity and anticipate the first half of capacity will be turned on by the end of April and the rest to be turned on by the end of June. The facility is fully contracted by Marathon for five years with a flat rate agreement. Our agreement begins upon energization.
Once fully energized, this location will bring Applied Digital to 280 megawatts of hosting capacity across all our facilities in North Dakota, all of which are contracted out to customers on multi-year terms. In addition to Ellendale, we're making great progress on our 200 megawatt Garden City facility in Texas. The construction of Garden City is complete and over 130 megawatts of miners are installed and ready to turn on. Our customers are continuing to send miners to the facility, and we are actively installing them. Given the unique behind-the-meter aspect of this facility, we are awaiting final approval on some final technical details and expect to resolve these issues in the coming weeks to begin energizing the facility.
Once we energize our Garden City facility, we anticipate it ramping faster than our Ellendale facility due to the amount of miners already installed. As a reminder, both our Ellendale and Garden City facilities are fully contracted with fixed prices. We are not exposed to any volatility in the cryptocurrency markets. Our 100 megawatt Jamestown facility continues to perform as expected and operated at full capacity throughout the quarter as we delivered revenue of $14.1 million in the quarter, exceeding the $12 million steady state capabilities of the facility that we previously discussed. As mentioned on our last call, we successfully retrofitted a small portion of our existing facility in Jamestown to accommodate HPC requirements to support a Web3 application with a non-crypto customer.
We have decided rather than to use the GPU capacity for the Web3 application, it would be better used for machine learning and AI applications and have onboarded multiple customers and recognized our first HPC revenue in the quarter. We also broke ground on a 5-megawatt standalone facility adjacent to our Jamestown site in December that will host several hundred graphics processing units for a machine learning application with a new customer. Our first GPUs in the new facility are expected to be operational later this month or early May. The build-out will be completed in two additional stages, with the first scheduled to come online this summer and the second later in the year. When finalized, we expect to have over 7,000 NVIDIA A100 class GPUs in the building, making it one of the largest GPU clusters of its kind in the world.
Importantly, it's worth noting that demand for hosting capacity across our facilities has not been impacted during the last several months, and we're exploring numerous opportunities. Let's discuss the HPC opportunity in front of us and why we're excited about the year ahead. While we continue to see robust demand from cryptocurrency miners, we aim to diversify our customer base and exposure to the growing segments of the HPC market as we believe that will be the highest return of capital in the long term for our shareholders. Our goal remains to get at least 10% of our revenue from HPC by the end of this calendar year and ultimately diversify our revenue to a 50/50 split by 2025.
We remain optimistic about the growth opportunities in HPC, which is expected to reach $900 billion globally by 2030 and remain well-positioned to capitalize on this opportunity. As data continues to grow at an exponential rate, more data centers will be required to store the data, and we believe our next-generation facilities are ideal hosting sites for HPC applications as they can accommodate the unique demands for this growing industry. Our data centers offer a more purpose-built solution, offering lower costs combined with higher computing power compared to traditional data centers that are typically focused on delivering low latency and high computing power.
We are well positioned for success in this space given our expertise in hosting Bitcoin mining and realize that HPC applications require a different type of engineering that more resembles what you would see in the basic world of Bitcoin mining because of its dense power. The density of racks is a key point as HPC applications require a different setup that provides sufficient power and cooling to handle those unique needs and properly scale efficiently. These applications don't require ultra-low latency. We believe that the deciding factor on whether these applications will be hosted comes down to the cost of compute. Our next-generation data centers are optimized for green computing. We aim to be the lowest cost compute provider with our access to renewable energy and air cooling.
To close, we remain confident that Applied Digital will continue to be a leader in digital infrastructure with our next-generation data centers. Demand for our services from both traditional customers and emerging HPC applications remains robust, which validates our position as a financially strong and leading digital infrastructure provider to serve various hosting needs. With that update, I'll pass it over to our CFO, David Rench, for a financial update.
Thanks, Wes. Good morning, everyone. Before I begin my remarks, I would like to note that like last quarter's call, since we did not have operations in a year-ago comparable period, we will not be providing any year-over-year comparisons. Revenues in the fiscal third quarter were $14.1 million, which were entirely attributable to our hosting operations. The Jamestown site operated at full capacity throughout the quarter. Cost of revenues in the fiscal third quarter were $10.5 million, consisting of $8.6 million of energy costs to generate our hosting revenues, $900,000 of depreciation and amortization expense, and $1 million of personnel expense for employees directly working at our Jamestown hosting facility.
Adjusted gross profit, a non-GAAP measure that excludes depreciation embedded in the cost of revenues and one-time electricity charges, was $4.4 million or 31% of revenue for the fiscal third quarter of 2023. Operating expenses for the fiscal third quarter of 2023 were $10.5 million, which includes $4.5 million of stock-based compensation, $3.9 million in other selling, general, and administrative costs, and $1.1 million in depreciation and amortization expenses. Net loss attributable to Applied Digital for the fiscal third quarter of 2023 was a loss of $7 million or a loss of $0.08 per basic and diluted share based on a weighted average share count during the quarter of approximately 94.1 million.
Adjusted net loss attributable to Applied Digital for the fiscal third quarter of 2023 was a loss of $7 million or a loss of $0.08 per basic and diluted share based on a weighted average share count during the quarter of approximately 94.1 million. Adjusted net loss attributable to Applied Digital, a non-GAAP measure for the fiscal third quarter of 2023, was a loss of $1.4 million or a loss of $0.01 per basic and diluted share based on a weighted average share count during the quarter of approximately 94.1 million. Adjusted EBITDA, a non-GAAP measure for the fiscal third quarter of 2023, was $900,000. Lastly, on our balance sheet, we ended the fiscal third quarter of 2023 with $22.9 million in cash and cash equivalents and $23.7 million in debt.
During the 3rd fiscal quarter of 2023, we received $11.7 million in net customer deposits and $32.3 million in net deferred revenue, which c ollectively amounted to a $44 million net cash inflow due to the structure of our commercial arrangements with our customers that incorporate upfront deposits and prepayments. In certain contracts, the prepayments are amortized back to the customers for the 1st year of their contract with no impact to revenue recognition, but the timing of cash flow with upfront cash to us is a major benefit for the company in that it helps with our CapEx funding needs as we build our data centers. Our balance sheet remains strong, and we have no exposure to Celsius Network, First Republic Bank, FTX, Signature Bank, Silicon Valley Bank, or Silvergate Capital Corporation. Now turning to guidance.
Similar to last quarter, we will not be providing explicit guidance for the forward quarter given the revenue materiality of our Garden City potentially coming online and the continued ramp of the Ellendale facility that should occur in the fiscal fourth quarter. Once both facilities are online, we will have nearly 500 MW of hosting capacity that we expect to put us in an annualized adjusted EBITDA run rate of approximately $100 million. That completes my financial summary. Now I'll turn the call over to Wes for closing remarks.
Thank you, David. I want to add a little more detail around our expectations for the current quarter. While we expect Garden City to energize during the quarter, if we exclude it completely and look at our expected ramp of Ellendale, we expect to generate approximately $24 million of revenue and approximately $4 million of EBITDA. Before we get to Q&A, I would like to quickly go over some key goals and initiatives for our company as we look to the future of Applied Digital. We remain focused on operating our Jamestown facility at high efficiency and look forward to energizing our Garden City facility and continuing to make progress at our Ellendale facility in the near term. We will also continue to build out our non-crypto use cases to demonstrate the broad capabilities of our next generation data center assets for HPC applications.
While the crypto industry remains volatile, we are well positioned to capitalize on strong demand for both crypto and non-crypto customers for our services. I remain optimistic about our future and want to thank all of our team members for their dedication and service to Applied Digital. With that, operator, let's open up the call for questions.
Thank you. If you'd like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star two if you'd like to remove your question from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. In the interest of time, we ask that you each keep to one question and one follow-up. Thank you. Our first question comes from the line of Lucas Pipes with B. Riley Securities. Please proceed with your question.
Thank you very much, operator, good morning, everyone. Congrats on the progress at Ellendale, I, I do want to ask about Garden City. You mentioned an expectation that it could come online in the coming weeks. I wonder if you could touch on the level of confidence on that for that statement. In terms of a faster pace of energization, which was also mentioned in the prepared remarks, is there ability to quantify the cadence of the ramp once you have those, you know, final metering issues resolved? Thank you very much.
Yeah. Thanks, Lucas. This is Wes. Let me try to explain the primary issue that we're dealing with. It is an issue with kind of behind-the-meter installation. We're not unique right now in Texas. There's several large projects turning on or trying to turn on. I think we're what might be the only Bitcoin mining project of its kind. There's several other kind of projects that are turning on. The issue is the way they settle on metering, and the visibility they want into the power gen plus the grid energy. There's an issue with double settlement or being double counted.
Or let's say there wasn't really a framework for this in ERCOT, and we're dealing with it mostly on the Oncor side from a metering perspective. ERCOT had a meeting last week where they had a rule change that should solve this issue. Now it's back in the hands of Oncor. We think that issue is solved, but they typically seem to move very slow. While it wouldn't be, you know, didn't happen the next day, you know, we expect some really meaningful progress on finalizing this. There's, just my opinion, some very simple solutions for it, but it just has had to go through the process, and it feels like the process is now largely complete.
With the rule change with ERCOT, it should solve the issue with Oncor , not just for us, but for several different people. And so we expect this to move forward in the very near term. So hopefully that makes sense. As far as the pace of energizing, recall Ellendale, you know, this is as we build, we're energizing. The buildings will go up much faster now. We had all the concrete poured. We're finalizing the buildings. All of them get worked on in stages. Now it's just turning on each building as it's completed, as the miners are racked, and as they're available to turn on. That's much more like we did in Jamestown. Because, you know, we originally expected Garden City to energize much earlier, the construction is complete.
The a large majority of the miners are already racked. We're not doing it in those types of stages, as we've done in North Dakota. When we have the final approval, when we're ready to turn on, all of these buildings are ready to turn on. You could see these come up in just a week or two. You know, it still takes a little bit of effort on each building, but they're already in place and ready to go. That's why we would expect that to ramp up significantly faster than we're seeing in Ellendale or that we saw in Jamestown last year.
That's very helpful. I appreciate the color, Wes. For my second question, I do wanna turn to HPC. Could you share some light on what an anchor agreement could look like in terms of revenue, unit economics, term, duration? Then, what would be the potential capital requirements for a large HPC kind of anchor tenant? Thank you very much.
Sure. Sure. There's a couple of different models that we're exploring with HPC. There's a model where, you know, the first ones that we're running in our, in what we call the HPC closet, it's the retrofitted portion of one of the mining buildings. We own those GPUs. We rent the capacity out. I think we mentioned we've had 3 customers live on those, you know, doing mostly machine learning, deep learning applications. It's been going really well. If you recall, we have a software partner that is helping us with that on the front end that loads onto the system. That's gone very well. As we move into the pilot phase of the new building, it'll be around 300 GPUs that we'll ramp up.
Those will also be owned by us. you know, we'll have customers, we have customers that are booked for those GPUs as well. the rest of the building can go in one of two ways. there's a potential for us to do just colocation on those areas. if you think about colocation, you know, a good area is we've gotten a kind of put a finer point on this. A good way to think about this is on a per megawatt basis, I'm just comping that to what we do with the Bitcoin mining because that's what we've done so far. You should expect more in the 10 to 12 times revenue versus per megawatt versus the Bitcoin hosting for the HPC applications.
Similar lift in EBITDA, maybe even slightly higher, maybe a little bit better margin on the EBITDA front from a HPC colocation. If we move forward with owning GPUs, if that is what we land with a customer, it's a significant step up again as far as revenue and EBITDA. If we own the GPUs in the entire facility, and we booked those out, and we would only do that if we had the right style of customer to do that and, you know, a very large, you know, solid counterparty to do that building would generate well north of $100 million of revenue and upwards of $50 million or $60 million of EBITDA.
We're still working through how we, you know, what the business model. I think it'll be a little bit split there. Right now, we're fully funded on our balance sheet now. You know, you saw where the quarter ended in cash. You know, as of yesterday, the company had about $41 million of cash, so actually a significant step up. That was the debt facility that went in place post the close of the quarter on the Ellendale facility that boosted that. We're fully funded to build the building and to put the GPUs that we plan to put inside the building currently. We're working on financing if we're going to move forward with owning more of the GPUs.
I think when you think about a longer-term model here, you should really think about a colocation model is what we're chasing, and then blend it in with some GPU ownership when necessary.
Very, very helpful. The, the $100 million revenue, what size of building would that be megawatt-wise?
That's the just the current build, the 5 megawatts. If we owned all the GPUs and did a leasing business, more of an integrated model like what you would see with an AWS, it would be. I said $100 million, it would be well north of $100 million.
For 5 megawatts. That's super helpful. Wes and team really appreciate the color, and all the best of luck. Thank you.
Thank you.
Thank you. Our next question comes from the line of Rob Brown with Lake Street Capital. Please proceed with your question.
Hi. Good morning. Wanted to get a little more color on the Jamestown facility. I think you said it ran at $14 million revenue above its sort of nameplate capacity. What drove that? Do you see that continuing?
Yeah. Good morning, Rob. Thanks for the question. Just higher efficiency in the quarter for us, was part of it. There was a little bit of a price lift. We do have the ability to increase price somewhat on our customers because we do have I think we've talked about this before. You know, last quarter, you saw a lower gross margin for us, and it's kind of the lagging effect of, you know, energy prices move around some at that facility, and then we, you know, do pricing changes for our customers, but it's lagging, and so we catch up with the pricing change. You know, you could see that go, you know, fluctuate up and down.
We are turning on a little bit more, so we have about 6 megawatts of Marathon containers at that facility that we expect to turn on any day that will boost the operations there a little bit. I, you know, I think that, you know, kind of that level that, you know, somewhere in the $13 million to, you know, on the upper bound, maybe $15 million or $16 million of revenue is reasonable on a quarterly basis from Jamestown.
Okay, great. Just wanted to get a little bit into the cash flow of the customer prepayments and deferred revenue. You said it would happen over 12 months, but how long does it take to sort of burn off that balance sheet prepayment and sort of normalize the cash flow against the EBITDA you're reporting?
Basically from the customers when their contract turns on, when we energize that portion of the contract, we begin to amortize that prepayment back to them. you know, it's four months. It's fairly simple. It's four months of prepayments amortized back over 12 months. Once we go through a full 12 months, that deferred revenue will be, you know, gone off the balance sheet. Obviously, from a P&L perspective, you get the deferred revenue, and you still get the same earnings impact.
From a cash flow perspective, you should, you know, think about as we're hitting that $100 million run rate for the first year of that run rate, you know, the cash conversion from EBITDA to cash flow is gonna be lower, but there still will be significant positive cash flow for the company as we amortize that back to the customers. When, you know, after you get through that year, the cash flow versus the EBITDA should move up more towards you know, 80% or 90%.
Great. Thank you. I'll turn it over.
Thanks.
Thank you. Our next question comes from the line of George Sutton with Craig-Hallum Capital Group. Please proceed with your question.
Thank you. Wes, one thing I wanted to just clarify, 'cause we had multiple clients concerned about the proposed permitting framework in Texas, the Texas Legislature. You are well beyond that. The modest delay we have had has nothing to do with that. That is a correct statement?
Yeah, that's correct. There's a couple of things going there, George, and it's new. Some of the stuff was just yesterday that, you know, we don't think. From that site specifically, we're beyond that portion. From that site specifically, you know, they're talking about, you know, the grid balancing or the demand response benefits in the legislature now. The setup of that site, it is not an area where we take advantage of those programs in Texas to reach the price on energy that we're contracting there. We don't believe it has any impact on us.
Gotcha. Great. On the use cases for the HPC, I wanna make sure I fully understood when we talked machine learning versus what I was viewing as large language models. I just wanna make sure, have we seen somewhat of a use case change as we look to deploy in new customers?
No, no. The large language models fall in that machine learning category, so it's, you know, it's training the model. When I say machine learning, it's, you know, training the model. If you wanna call it AI or machine learning, but the applications that are running now include the what we call natural language processing.
Okay. Perfect. Thank you.
Thanks, George.
Thank you. Our next question comes from the line of John Todaro with Needham & Company. Please proceed with your question.
Thanks for taking my question, guys. Two questions here. First, on the Bitcoin mining piece, can you just remind us of, for at least your major mining contracts, when those are up for renewal? Then any expectations you have with the halving coming up in Q1 2024, which should raise those mining costs quite significantly.
Yeah. The contracts are, you know, the vast majority is contracted under 5-year contracts that are effectively just starting as we energize, excuse me, as we energize Ellendale. In Jamestown, I think off the top of my head, the exact contract life remaining, but it's probably in the 2.5-year timeframe. That's the majority of our contracts. On the halving, yeah, the halving event, I mean, it's pretty simple to go, you know, to think through that. We think our current customers, you know, I think we run a really efficient operation, and our current customers are very profitable at the current Bitcoin price and significantly below the Bitcoin price when you think about just about variable cost.
effectively, you know, the cost will increase, double at the halving event, you know. It really depends on two things. it's, you know, the price of Bitcoin and then the network hash rate. you know, John, the thing about Bitcoin over, you know, the life that it has had is generally there's a self-correcting mechanism, you know, in the network of hash rate versus price that allows miners to stay profitable. There will be periods of time where they won't be. at the end of the day, you're looking for the most efficient guys, I think are the ones who, you know, will stay online post the halving.
We're running only S19 Pro, and the majority of what we're turning on now and will be at the new facilities are all XPs. We're running all the latest model mining equipment, and we think that we and our customers are well positioned for that.
Got it. Thanks for that. Just the other question on the HPC segment. Obviously a lot of interest in the space recently. I would imagine that's kind of increased some of the competition there. Could you just discuss that a little bit and some of the competitive landscape there?
Yeah. I expect there to be, you know, plenty of competition and people entering the space. I expect it from the traditional data center providers. I, you know, expect new entrants to come into the space. You know, I think it's a really large opportunity. Just to kind of frame this opportunity and why I think we have this opportunity and what the opportunity set is, you know, we talked about these next generation data centers a fair amount. Let me just give you know, kind of some comps of, you know, traditional data centers versus what we're building and what we see in the future. You know, traditional data center we talked about, it's generally built for ultra-low latency, you know, interconnect, video streaming, all of those types of things.
Traditional data center built, you know, racked, you know, power to the rack. Typically for a full rack, they deliver about 7.5 kilowatts, and generally their capability is somewhere around 10-15 kilowatts to a rack. When we load a rack full of, you know, an NVIDIA box with eight A100 GPUs, which is the, you know, kind of the gold standard right now for machine learning and AI, that rack, if you want that rack full, it requires about 40 kilowatts to power that rack. You're looking versus what traditionally is there, almost a fivefold, you know, increase in the power needed, and then that, you know, really directly correlates to the amount of heat created.
You know, we're talking to the largest players in the industry, and we're spending a lot of time with them. You know, the view is things are going from 40, you know, to 70 to 100 kW needs per rack. You're bringing this power density, you know, it's increasing significantly, and it's really difficult to retrofit older data centers to do this. Then they might. You know, then you're thinking about the amount of power that they need. I think you're gonna see a trend for specifically machine learning and AI because it's a very unique load that looks a lot more like what we're handling now. To move them closer to the power source, you don't need this ultra-low latency aspect for these. We're positioned very well.
We have, you know, large amounts of low-cost power. We're sitting mostly in very cold locations. You know, and our data center, for example, you know, the airflow is significantly higher than you would see from a traditional, just because we want to use air cooling to cut down on the electricity usage and lower our cost significantly. A traditional data center, you know, maybe half a mile per hour of airflow through. You'll see our new facility is designed for around 8 miles per hour of airflow, just massive amounts of airflow for the cooling, because you know, the climate in North Dakota is absolutely perfect for this. We're sitting on a fiber grid. We have 100 gig connectivity now, going to 400 gig at the end of the year.
We have really good fiber connectivity, which is also necessary. You just don't need the ultra-low latency aspect. We're sitting in a really good spot. Is it easy for everyone to go into this space? You know, just like we've seen in building out hosting capacity for Bitcoin mining, that was hard. I think this is even harder. We've put a really good team in place. You know, we pulled some people out of a large company that builds data centers here in the U.S. and actually internationally as well. You know, they were building data centers for some of the largest hyperscalers. We have people that are experienced with doing this. We've spent a lot of time designing this specific design. Now we're implementing it.
I do think we're, you know, we have a significant lead, you know, versus almost everyone out there. We've seen a few other smaller players that have been doing this for a few years. You know, I think you're gonna see some of the larger ones move towards it. I do think we have a, you know, a pretty good advantage, given where we are now, where the sites that we already have, the operations that we already have, and the knowledge base we have with our, you know, really on our electrical engineering as far as being able to deal with this kind of power density already. I do expect competition.
I expect it to be an extraordinarily large industry, and there's no way you're gonna have one industry like that without a lot of competition.
Yep. Got it. That's great. Thanks, Wes. Appreciate it.
Yep. Operator?
Thank you. The next question is coming from Mike Grondahl of Northland Securities. Please go ahead. Hey, guys. Thanks. Wes, when do you think you kind of have a green light on the HPC opportunity? Is that when the 5 megawatts are sort of up and running and that's shown everybody you can do this? Is that this summer, this fall? When do you think you have that? What could the HPC business look like in 2-3 years in terms of number of megawatts and sort of, you know, a rough range for revenue?
Sure. I gave the kind of the revenue comparison earlier. I can, I can repeat that if you want me to, but just to kind of frame where we think, you know, HPC is as far as, you know, revenue versus the Bitcoin mining. The, the green light, we've kind of we're definitely at maybe like a, we've moved from red to yellow already. We're operating in our site. We see that these work. We, you know, we see the network connectivity works. We see the software works.
We see this setup works. Now, you know, as we bring the building on, there will be some, you know, some fine-tuning on the, on the five-megawatt building for, from a design perspective for sure, just like we did, you know, some tuning up from when we built in Jamestown versus what we built in Garden City and Ellendale. We're really close, and, you know, the conversations that we're having and the people we're having these conversations with, has really, you know, given me and the team a large amount of confidence that we're, you know, moving in the right direction.
As far as, you know, what that could look like right now, you know, trying to think about what that could look like a few years out, we talked about getting, you know, to a 50/50 split. I think the demand that is out there for this type of data center build is large right now. You know, the people that we're talking to about hosting or about GPU rental, it's very large quantities. You know, it's something that, you know, depending on how a few of these things go over the next few months, this business will be, you know, potentially very material, you know, by the end of this year. It could ramp extraordinarily quickly.
If it starts, Mike, if it, if it starts to go this way, you know, it's kind of one of these where I think there it'll be a really long runway of growth. The conversations we're having right now, you know, I You know, we're seeing demand, you know, of, you know, people wanting individual demand of, you know, of 10 or 20 megawatts and some, you know, 50 and 100 megawatts of HPC build, which is a very large amount, given, like I said, what the, what the numbers were that I gave you earlier in the call about the revenue comparison for HPC versus Bitcoin mining. We're seeing a lot of demand.
You know, we're getting to some certifications that we need that we'll get in the summer. I don't know that that's gonna stop us from signing some large customers prior to even reaching that because you know, these are really in the stage of development projects that will need to be built and turned on. The cadence might call for some of these to come on or to be signed at least earlier than what maybe I initially expected.
Got it. Okay. Hey, I took away as the revenue co-comparison, it's like 10-12 times-
Yeah. For revenue.
What revenue?
That's right.
-you generate at the miner. Okay. Got it.
Yeah.
Thanks.
That's for Mike, that's for colocation. If we just do the hosting, that's a, I think, a good comparison right now. If we own the GPUs, it's significantly higher than that.
Got it. Okay. Thanks.
Thank you. Ladies and gentlemen, as a reminder, if you'd like to join the question queue, please press star one on your telephone keypad. Our next question comes from line of Kevin Dede with H.C. Wainwright. Please proceed with your question.
Hi, Wes. Thanks for the discussion on, you know, on the HPC side. I guess it sort of begs the question, and given that you seem pretty fully contracted out on 500, what other... Like, can you give us a, I guess, a peek into the pipeline for your next round of builds, your next site acquisition, and that kind of timing?
We have a very good site identified in another cold region, not in North Dakota, that has a good power price for us. We're optioning that site. I actually, you know, we have demand for Bitcoin mining for that site, we're seeing the amount of demand I think we're seeing emerge on the HPC front. I think it's, you know, the likelihood that it goes towards HPC is extraordinarily high. We do have another site identified, and we also expect to expand at the Jamestown site even further. Those are the two primary ones in the pipeline.
Okay. Does Jamestown entail like a, another set of buildings, or are you gonna kinda do it in those 5-megawatt increments?
The next build for HPC in Jamestown would be a larger above the 5 megawatts. Recall, I guess, I recall, but originally in Jamestown, we expected to go to 200 megawatts. I think we can push that beyond the 100, you know, with the power provider there, with HPC. I'm not certain what that limit is for us just yet. We do think that we can expand at that site. We'll push more towards new sites. We have identified one very attractive site that we expect to move forward with.
Okay. Can you just dig in a little bit on the colocation versus, I guess, the machine-owned business models in HPC, Wes? How would you approach that from a financing perspective?
From a financing perspective, you know, the goal. Owning the GPUs would be significantly more expensive. You guys probably know the pricing on these types of GPUs. They're extremely high. You know, our team on the finance side has done a great job to date, you know, getting us very low-cost financing for sites that we have for that particular market, if you wanna own the GPUs, there is a much better, really well-developed financing mechanism and financing market for that. We, I, you know, we've already are down that path, and I think we'll be able to tap some of that financing, and we are on kind of the initial build.
You know, if it is gonna be extraordinarily large, you know, we're working on solutions around that. It is capital-intensive. If you, if you wanna do the full, you know, the fully integrated model is one that we will pursue some. To what amount, I don't know, and it's gonna be, you know, could it be limited by our ability to finance that? Yes. On the colocation side, I don't think we'll have any issue financing that for HPC builds for colocation. The customer set there that we'll contract for that, you know, it's large companies, you know, triple-A type credit companies that I don't think will...
With our ability that we've had to finance the Bitcoin mining facilities once contracted with customers, I think it's gonna be significantly easier, even in a much more difficult financing environment that we're seeing right now, just because of the customer profile for those types of colocation buildings.
Okay, would you mind just sort of touching on the JV that you sort of set up on sort of picking up excess rigs, and how's that looking? What are you thinking about that now? What can you speak to?
Yeah. We, we didn't pull any significant amount of money in, unfortunately, on that, on that JV. We saw the opportunity there, we just didn't get a lot of demand on the... You know, it was basically, Kevin, that was set up, you know, as like a fund structure, like an SPV, to go out and do that. The, the opportunity is kind of, I wouldn't say totally come and went, but, you know, the price of the, of the equipment, you know, has moved up materially. That would've been a great trade, had we, had we been successful in pulling money in. At the time that we were doing that, you know, we had a starter, but we never got critical mass, to push that forward.
There just wasn't a huge appetite on kind of the investor front to put money into the space, which maybe was a signal that we're, you know. It was a good opportunity, but we never really pulled anything material into that.
All right. Thanks so much, Wes. Good effort. I appreciate the insight.
Thanks, Kevin.
Thank you. Ladies and gentlemen, that concludes our question and answer session. I'll turn the floor back to Mr. Cummins for any final comments.
Thanks, operator. Just to clarify in the prepared remarks, David Rench repeated the adjusted earnings number twice at 2 different numbers. The appropriate one is a loss of $1.4 million and a loss of $0.01 on the adjusted net income. Outside of that, I just wanna thank the shareholders that are on the call for their support and thank our employees again. It was an extraordinarily difficult winter in North Dakota, and it's still going with more snow recently falling there as well. Everyone's done a great job getting the buildings up, operating the buildings, and I really appreciate everyone's effort on that, and we'll talk to you next quarter. Thank you.
Thank you. This concludes today's conference. You may disconnect your lines at this time. Thank you for your participation.