Good afternoon, and welcome to Applied Digital's Fiscal Fourth Quarter 2024 Conference Call. My name is Shamali, and I will be your operator today. Before this call, Applied Digital issued its financial results for the fiscal fourth quarter ended May 31, 2024, in our press release, a copy of which will be furnished in a report on a Form 8-K filed with the SEC and will be available in 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 the remarks, we will open the call for questions. Before we begin, Matt Glover from Gateway Group will make a brief introductory statement. Mr. Glover, please proceed.

Great. Thank you, Shamali. Good afternoon, everyone, and welcome to Applied Digital's Fourth Quarter 2024 Conference Call. Before management begins their formal remarks, we'd like to remind everyone that some statements we're making today may be considered forward-looking 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 to differ materially from those described in the forward-looking statements. For more detailed risks, uncertainties, assumptions relating to our forward-looking filings made with the Securities and Exchange Commission. We disclaim any obligation or any undertaking to update forward-looking statements to reflect circumstances or events that occur after the date except as required by law. We will also discuss non-GAAP financial metrics and encourage you to read our disclosures in the reconciliation tables, the applicable GAAP measures in our earnings release, which can be found on the Investor Relations section of our website carefully as you consider these metrics. We refer you to our filings with the SEC for detailed disclosures and descriptions for 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, our quarterly report on Form 10-Q. You may get Applied Digital Security and Exchange Commission filings for free by visiting the SEC website at www.sec.gov. I would like to remind everyone that this call is being recorded and will be made available for replay via a link available in the Investor Relations section of Applied Digital's website. Now I'd like to turn the call over to Applied Digital's Chairman and CEO, Wes Cummins. Wes?

Thanks, Matt, and good afternoon, everyone. Thank you for joining our fiscal fourth quarter 2024 conference call. I want to start by expressing gratitude to our employees for their ongoing hard work and service and supporting our mission of providing purpose-built infrastructure to the rapidly growing high-performance computing industry. Before turning the call over to our CFO, David Rench, for a detailed review of our financial results, I'd like to share some recent developments across our business. During the quarter, we faced several challenges that impacted our financial performance due to the facility power outages in our data center hosting business. Despite these short-term setbacks, the company has made significant progress with our key growth initiatives, including the development of our cloud services business and the construction of our purpose-built 100-megawatt HPC data center in Ellendale. As previously announced, we executed an LOI with a U.S.-based hyperscaler for 400 megawatts at our Ellendale campus, inclusive of our current 100-megawatt facility and 2 forthcoming buildings. Now I will provide an update on each of our business units. Let's begin by discussing our data center hosting business. Our 106-megawatt Jamestown facility has consistently met expectations, operating at full capacity with uninterrupted uptime throughout the quarter. This achievement marks the seventh consecutive quarter of full capacity operation for this facility. While we are pleased with Jamestown's performance, we encounter challenges at other facilities. As previously disclosed, our 180-megawatt Ellendale facility in North Dakota experienced a power outage starting in January, which we determined was caused by transformer failures. By the end of the fourth quarter, we had successfully replaced the transformers and related components with equipment from industry-leading North American manufacturers. As a result of the transformers being replaced, we now have 286 megawatts of data center hosting capacity for our blockchain clients…

Thanks, Wes, and good afternoon, everyone. Let me begin by addressing the complexity of this quarter's financial reporting. We reported an adjusted EBITDA of approximately $4.8 million. However, several onetime items significantly impacted our financial performance and comparability to prior quarters. Notably, a large portion of the power was out in our data center hosting facility in Ellendale, North Dakota, which then largely came online right at the end of the quarter. Additionally, we incurred many onetime professional service expenses primarily related to our capital-raising initiatives, financial analysis for data center financing and strategic transactions. We continue to pursue all available remedies to recoup lost revenues and additional costs incurred from the transformer outages. Let's delve into the results of the quarter. Revenues for the fiscal fourth quarter of 2024 were $43.7 million compared to $22 million for the same period in 2023. The increase was primarily driven by expanded capacity across our data center hosting facilities and revenue contributions from the cloud service contracts. Specifically, our data center hosting segment generated $26.9 million in revenue, while our Cloud Services segment contributed $16.8 million. Turning to cost of revenue for the fiscal fourth quarter of 2024, it amounted to $46.3 million, up from $15.9 million in the same quarter of 2023. This increase can be attributed to higher energy costs resulting from the increased number of megawatts used to generate hosting revenues. Additionally, depreciation and amortization expenses, along with personnel costs rose due to the growth of the business as more facilities came online. Selling and general administrative expenses for the fiscal fourth quarter of 2024 were $31.3 million compared to $12.3 million in the prior year's comparable period. The increase was primarily due to start-up costs as we ramped up the cloud service business. This included higher depreciation, amortization…

Thank you, David. We understand the past year has seen challenges as the company has faced a multi-month outage at our Ellendale hosting facility and the financial burden of funding $1 billion data center while funding our cloud business. Nevertheless, we've made substantial progress towards securing greater than a decade long lease agreement with a Fortune 50 company for our Ellendale campus. Finalizing this lease will solidify our position as a leader in the HPC data center market and have a cascading effect on our ability to secure asset-level financing and to develop additional sites. Throughout this process, we've diligently worked to secure project-level debt for our facility with multiple interested parties. Upon execution of the lease with the hyperscaler, we anticipate our financing partner, CIM Group will release additional capital to support ongoing construction while we continue to work toward project level finance. In summary, despite significant challenges this quarter, largely due to external factors, we remain fully committed to delivering strong long-term shareholder value. Our vision is to become a development platform capable of building and operating multiple HPC data centers. This starts with our Ellendale campus and continues with 3 additional campuses we are actively marketing today totaling 1.4 gigawatts. To support this vision, we have added several industry veterans to our team, and we're already working on the design of our next 2 buildings, which will provide 300 megawatts of capacity. We are incredibly proud of the progress made this quarter and look forward to providing further updates as we move into fiscal 2025. Looking ahead, we believe Q4 marked the bottom for our revenues and anticipate sequential improvements in the top line as we enter the first quarter. We now welcome your questions. Operator?

[Operator Instructions] Our first question comes from the line of Lucas Pipes with B. Riley Securities.

Wes, to you, my first question is on the Ellendale HPC campus. And I wondered if you could speak to the percentage completion on the first 100 megawatts and the remaining capital requirements from here?

Sure, Lucas. So we have a little over $200 million into the facility as it stands today. The building, I'm sure you've seen some of the video footage is fully enclosed, we're progressing with MEP at this point, and then we'll do the final fit-out later this year, early next year. And we go back to -- so this facility will be roughly $1 billion -- a little over $1 billion for the first 100 megawatts, which is on par with standard Tier 3 data center capacity. And so with the $200 million in, I think I've talked about this in the past. We're working with banks that typically provide project-level finance in this space. We have term sheets and we've actually selected a bank that we're moving forward with. And the expectation is that they'll fund the remainder in the -- we've seen 80% to 90% loan-to-cost quotations from this, we just need to get past the lease finalization and the project-level finance process is already running at this point. So we'd hope to close that shortly thereafter.

Wes, did I hear it right that you said kind of fit out early next year? And would that include completion? Or how would you frame that up in terms of that time line?

So Lucas, there's a couple of different timelines that we work through here as we've gone through the process. There's our build timeline, and then there's when our customer wants the facility ready for service. That's not completely locked in at this point. But our schedule from the build perspective is what we've talked about in the past being ready late this year or early next year.

Change in topics really quickly. On the cloud services business, can you speak to the ramp in your fiscal first quarter, how many clusters did you have online? And how would you kind of frame up the revenue opportunity Q1 -- fiscal Q1 here versus just this fiscal Q4 period that you just reported?

Sure, Lucas. So where we are there, as we said in the prepared remarks, we're 6 clusters online in Q1, and it won't be for the full Q1, but those came online, I believe, in June, so it will be for the majority of it. So with the 6 clusters up, that business is roughly $110 million revenue run rate on an annual basis. And right now, I've talked about on previous calls that our team is very focused on the enterprise market. We started with the AI labs and the AI start-ups, we're very focused on the enterprise market at this point. We have been working on it for 7 or 8 months now. And we've probably seen recently, we've hired a new CRO that came from IBM that's very focused on that market. And so we see the demand there, and we'll continue to ramp that. The other piece of the puzzle here is we have been working on financing that we're putting in place that it finances this deployment, this equipment in a way that makes it run through our income statement more fairly versus what it does today. So we don't want to keep adding a lot with the capital lease structure that we currently have just because of -- you see what it does to the financials because we're forced to do the depreciation over 2 years versus the industry standard of 5 or 6, so those two pieces are coming together, and then -- but we do still see a significant amount of demand, but really growing demand from the enterprise side of the market.

Our next question comes from the line of Rob Brown with Lake Street Capital.

Just following up on AI cloud -- on the AI cloud business, I think the 6 clusters, what's sort of the range of clusters, do you think, you can grow that business into? And how do you think about that versus the data center business in terms of capital investment.

On both of these businesses, Rob, we think about these businesses in the same way. These are asset-heavy businesses, capital-intensive businesses. We need to have the right mechanisms in place to do asset level financing for both of them. And so we've talked about that on the data center side. We're almost there. We talked about it pretty extensively on the GPU side. I think we're almost there as well. And so when we have the correct asset level financing mechanisms in place, I think we'll be able to grow significantly on both sides of that business. On the AI cloud portion of the business with the data center capacity that we have in place today and what's available to us as we go through '25, we can generate significant growth in that business, and we see the demand for it, but we need to get the right capital base in that business to grow it aggressively. Right now, we're kind of feeding both of them, and we need to move that financing down to an asset level from a company level.

Okay. Great. And then on your hyperscaler contract, I assume the -- just in terms of the timeline, the long lead items there are the due diligence, I assume, in the site and the technical due diligence, but sort of what percentage of completion are you at in terms of getting the contract through its process?

It's hard to handicap that, but I would say we're north of 90% of the way there.

Our next question comes from the line of Darren Aftahi with ROTH Capital Partners, LLC.

Just following up on the lease. I guess from a big picture perspective, like how much of this is small details being ironed out versus a larger framework, and I guess my other question in regards to the lease is just how much is a backlog in the law firm that actually doing this [indiscernible] this process.

That's a good question, Darren. So as far as the details, there's large items, and then there's very small items and sometimes the small items can take as long as the large items there. But there is an extensive process around fiber connectivity, power, firm power, redundancy on power that took a significant amount of time to complete, but there's a lot of small details. So I would say it's a split between those, and it's funny, sometimes some of the smaller details take longer. But the question on the -- it's hard to get a view on the law firm, but there's only a few that deal with most of these contracts. And so there definitely is a backlog out there, but I can't handicap or just talk about how much slowdown is due to that. This is our first time through. So this is the only environment we've experienced.

Got it. And just one more on your pipeline of other campuses. I guess, one, where are you in the marketing process? And then two, what's your propensity to diversify your customer base with those other campuses versus, say, if the existing U.S. hyperscaler that you have the LOI will have more interest in capacity if they wanted to take that down, just if you kind of compare and contrast those thoughts.

Sure. So we've just recently kicked that process off, there's -- we send out tearsheets and binders for these sites. And we have a more refined process, I would say, at this point than we did for the Ellendale campus, but we're still pretty focused on hyperscalers for these sites. We're focusing on sites that are at least 200-plus megawatts of critical IT capacity. So that is going to equate to depending on what part of the country you're in, somewhere between 250 to 300 megawatts of utility power, so those are the types of sites and customers that we're focused on. There's other customers in the market, but we have a pretty narrow focus of how we go to market with these new sites.

Our next question comes from the line of Mike Grondahl with Northland Securities.

Wes, could you repeat what you said, the 6 clusters are driving an annualized run rate of how much revenue, I heard $100 million, but it kind of came through a little garbled.

Yes, it's about $100 million to $110 million, Mike.

Okay. $100 million to $110 million. And I think what you were saying is, hey, there's a lot of demand for that business, but you're kind of going to stay at about 6 clusters until you get a better financing structure in place, and then you would expect to see more growth. Is that a fair summary?

Yes. That's a good summary is we're working on a better financing structure because you see -- I mean you can look at kind of where our adjusted EBITDA is versus where our earnings are, and that's driven largely by the accelerated depreciation schedule on those GPUs. And so we need to have the type of financing, one, that is at the right cost for us. And then two, allows us to depreciate these over the useful life rather than during the financing period, if that makes sense.

Sure, sure. And in the press release, you kind of referred to this development platform on the HPC side with the potential to go from 400 megawatts to 1.4 gigawatts, do you think that first hyperscaler customer is interested in more than the existing 400 you're talking to them about? Could you end up being a one-stop shop for them, for another big project or two?

I definitely think we could, but there's other demand in the market as well from similar type customers. There's a significant amount of demand for this capacity, especially near-term power. So '25 power -- 2025 power is done unless we have that for Ellendale. Now we're focused on 2026 power and then 2027 power. But I think near-term power is key, and you need to have the right type of power. Again, one of the reasons when we talked about selling our Texas facility, we don't believe that facility works for HPC, at least not the way we want to develop it. So you need the right type of power. So we're really focused on near-term power. And Mike, just to clarify, the 1.4 gigawatts is just the additional 3 sites. It doesn't include Ellendale. So that's what we're focused on, the market there. But I think the potential current customer as well as others like them, there's a significant amount of demand for near-term power in the market.

Our next question come from the line of George Sutton with Craig-Hallum.

Wes, nothing on the call has surprised me except for one statement. You mentioned it's not clear when the customer might want to go live with the lease. Can you just clarify that statement? My assumption has been we want to get this up and running as fast as humanly possible, and that's why you're building out as rapidly as you are.

Yes. So that's correct, George, a correct way to look at it. But there's different stages for this as far as when you're doing fit-out, what we're adding in for the fit-out. So there's a lot of moving parts there. I mean it's fairly buttoned down, but we'll save the details of that for when this is finished, and we can share that publicly.

Okay. And relative to the GPU financing structure, our understanding, knowing that market pretty well is it's gotten a lot better. It's gotten a lot more financeable, can you just talk about the level of interest you've had from folks in putting a structure together there? Kind of where does that stand?

Yes. So we started this process a while ago, and you are correct in that just the interest level has grown and the cost of capital has come down. It really, George, depends on the kind of the offtaker here, just like on the data center side. So the financing pricing depends a lot on the offtaker. And so your highest costs are going to be AI labs and start-ups and then enterprise is going to be significantly lower. And then if you were doing the deployment with a hyperscaler, it's going to go even lower from there. But we're seeing a significant amount of new entrants into that market, and we've seen a lot of interest, and I think we'll have been successful in putting that structure in place for ourselves.

So just to clarify on that, the financiers are saying we're happy to put them together. We want to see the customer names. And once you provide a high-quality enough name, we would want to fund this. Is that chicken-and-egg to some extent.

It kind of is, George, what you should think of is a facility that gets put in place, and it's a drawdown facility as you buy equipment and you do show customer contracts. And typically, the kind of the structures that we see in this is you'll be feel -- form like a bankruptcy remote SPV and then the loans go into that, the GPUs go into that and the customer contracts go into that. And in some cases, the colocation contracts go into that as well. And so it's kind of a self-contained financing solution.

Our next question comes from the line of John Todaro with Needham & Company.

Wes, two for you. First, just trying to due diligence as much as possible, kind of the timeline for that lease agreement. So you had mentioned they finalized their technical requirements, is that kind of you think about the first stage and how many stages would you say there are? Or is it really now just kind of down to some final details. And then second question, you had mentioned the Garden City site that you didn't think it was suitable for HPC. Obviously, a lot of the Bitcoin miners out there are kind of trying to go down that West Texas route. It hasn't been done yet. Just curious why did you think that, that site and maybe that region, would it really be possible for HPC.

So John, on the -- I would not say, I mean, it's all kind of -- when you say is that the first step of the process on the lease side, all kind of goes together. But I would say we're finalizing the last details of that is the way I would characterize it. And then on what we've learned through this process and what we had seen previously but especially through this process is that I don't think, for us, for our company, everyone else has their own strategy. But I don't think for us, having a large flexible load in the ERCOT market is going to be suitable for the kind of customers that we are pursuing in that business. I don't know what changes will happen, but I know there are changes in the legislature that have been proposed in the legislature in the state of Texas that will make even harder with the large flexible load structure to serve, again, the type of customers we're looking at serving. So right now, we're not pursuing anything in the state of Texas at the moment. That could change in the future. But there's firm power, power redundancy and not just redundancy of what you build on the site, like UPS or backup gen, redundancy in having multiple power lines or feeds and from different directions into the substation or switching station that you're drawing power from. There's a lot of pieces that we've learned going through this process, and that's the type of sites that we have locked up and are locking up in marketing, but there's -- I'm not that bullish on just specifically the large flexible load in the state of Texas.

And our next question comes from the line of Kevin Dede with H.C. Wainwright.

I was wondering if you could take a little time just to characterize the overall environment. Understand you're targeting 1.4 gigawatts. What's the competitive environment like? And how are you confident that Applied wins versus your competitors.

Sure. So there's two parts to that. So from the environment, as I mentioned earlier, we see very strong demand for near-term power. So we're very focused on power that can come online through the end of '27. So power is the #1 bottleneck at the moment. And then when you step back to what I believe is the #2 bottleneck, it's supply chain. And there are a lot of stretched components in the supply chain. I'm happy to walk through those, but that's the second part. So we have a significant amount of experience in the supply chain. A lot of the things we -- for the first data center build we had ordered last year, we bought our places in line for switchgear transformers, a lot of the key components that are needed to build these. So those are the 2 pieces. And I think we're in a great position from both of those kind of competitive dynamics. I think the biggest thing for us is getting through this first lease that will be something that validates what we're doing as a company in North Dakota and our ability to go and find these sites, develop them and bring them online for very large companies and demanding companies. And so there's a lot of competition in the market. We see the majority of our competition being private data center companies that have been building for hyperscalers for years already, decade even. So those are the ones that we see as the majority of the competition in the market. And there's a handful of those platforms out there that are building hundreds of megawatts or in some cases, over a gigawatt of power out for hyperscalers. So this is not necessarily a new market. It's just a new style of data center.

Well, as you go up to bid for a site and you see these other private companies that have been doing this a little bit longer than Applied has, why would you win that contract? Why should we believe that you're going to have a foot up in that competition?

So I think it's important to understand that the companies that we have historically -- that have historically been in this market and the way kind of data center the industry has been historically is when a company wants to build new capacity, they typically look for the region they want to be in. So let's say, I'm sitting in Dallas. Let's say they want to be in Dallas. And so there's a couple of different tools that you can license for locating data center space in the DFW area. And those tools really focus on the fiber maps. And so what it will do is basically match the type of fiber connectivity you need and then available real estate on the map, and then you'll find a location, purchase that location and then you'll make a request to the utility for power. And so what's happened in the market now is companies are still doing that, but those power requests goes in a queue and in a lot of the most popular markets that is stretching out 5-plus and in some cases, over 10 years before you're going to get that power, especially at the amounts of power that are needed for these new style facilities. And so that's the big gating item. Now us on the other hand, we go about it kind of the inverse way. We find power availability and then we make sure that, that powers the type of power that can be used for our type of development. So what I've mentioned before on redundancy. And then we make sure that the fiber connectivity is there. And now we've refined that even further to include kind of latencies to different locations. And then we can go and either start developing that site like we did in Ellendale or marketing some of our other locations. So I think we're unique in one of the few companies that goes about it that way and is this far along, if you look back at our history, we started down this path in 2022. And so we have a few years of experience doing this, and now our knowledge level has went up dramatically in the last 6 months inside the company about how to go about this. So I think just the availability of power and the right kind of power at the right type of sites with the right type of fiber connectivity or what really gives us an advantage.

Would you think it's also fair to say that you're looking under rocks and perhaps more rural locations with less concern on latency.

So we are, for sure, that's what I was mentioning earlier about how traditional data center companies go about it, and we go about it in a very different way. We've definitely tightened the bands on what we know about kind of the latency requirements. So we're not necessarily going to be looking at really far front locations, but we have a lot more knowledge base around the type of latency requirements in North America that we will need to meet for the site. So it's helped us with our site selection even more.

Can you talk a little bit about the where Applied stands now on the cloud AI biz. How many sites do you have, and what's active? And what do you think once you're financing -- your structured financing in place, how you'll be able to access those this fiscal year?

Yes. So we have a lot of the same locations, right? We have Minnesota. We have our site in Jamestown, we have colo's capacity in Denver and Salt Lake City. Those are the primary large ones. We have a smaller one in Las Vegas. So we have that capacity. There's open capacity, which is, I think, a very big asset in the market. It's extremely difficult to find and ready for deployment. We have demand that's coming through on the enterprise side, and there still is strong demand in that market. When we get the right type of financing in place, we'll be able to reaccelerate the growth of that business.

So all told, can you give us a rough ballpark on the megawatts that you have access to?

We have megawatt capacity secured of just over -- for deployment just over 30 megawatts. Then Kevin, that's ready now. That's not for in the future.

Yes, that was the nature of my question. I appreciate the color, I appreciate the color.

Thank you very much. Thank you. And we have reached the end of the question-and-answer session. Therefore, I'll turn the call back over to Wes Cummins for closing remarks.

Thanks, operator. Thanks, everyone, for joining the call, and look forward to speaking to you next quarter.

And this concludes today's conference, and you may disconnect your lines at this time. Thank you for your participation.