Hyperscale Data, Inc. ($GPUS)

Good afternoon, everybody. My name is Todd Ault. I'm the Executive Chairman and Founder of Hyperscale Data. I see a lot of people on here that are familiar with the company, some that are new to the company. I've been the Executive Chairman, I believe, since 2018, where we laid out a 10-year vision for the company. On the call with me is the CEO, Will Horne. Will Horne and I have been business partners for 20 years on and off, mainly pretty much on. And I see a lot of people saying, Todd is the CEO. And I'm sure there are days where Will would prefer I was the CEO, but Will and I run the company together. Will is an operational genius and really a financial genius too, and a great partner. So welcome to the conference call. It's the first one we've had in a long time. Will is going to read an abbreviated forward-looking statement. Gary, can you please switch the slide? Thank you.

So before we begin, I'd like to remind everyone that today's discussion includes forward-looking statements within the meaning of federal securities laws. These statements are based on current expectations and assumptions and involve risks and uncertainties that may cause actual results to differ materially. Please refer to our filings with the SEC, including our most recent annual report on Form 10-K that was filed on April 15 as well as a copy of this presentation that was filed on Form 8-K today. We undertake no obligation to update forward-looking statements, except as required by law.

Thanks, everybody, and thank you, Will, for this. I do want to address a couple of things as I've been bombarded with questions. Today's call will not deal with Ault Capital Group, which is inside Hyperscale. We are an unusual company, and that is we have, in my opinion, a very small market cap with a company that has forecasted almost $200 million in revenue for 2026 and has various subsidiaries in foreign countries, including Israel, the United Kingdom. We have an office in Korea, and we have offices throughout the United States in multiple businesses, including cranes, hotels, defense, the lender and the blockchain project, which is in 172 countries. So it's mildly difficult as a small public holding company. And as we've disclosed multiple times, the company is planning to break up into 2 companies: Hyperscale Data, which will remain public on the New York Stock Exchange American and then Ault Capital Group, which we hope will list on the New York American. And with that in mind, we'll continue. Now Gary, can you please switch the slide? Thank you. So we have a fundamental thesis here, and let me read it to you a little bit, and that is I want to start to talk about the fundamentals here of the economy is not going to be priced the way the past economy was. In the next decade, we believe that compute power will become currency. If you control compute, you control intelligence. If you control power, you control compute. If you control both, you're at the center of the next economic cycle. This is the thesis that's driving Will and I and the entire company and the separation of the company. Now let me address a few things directly. Some of the questions about potentially selling data center. Power is very valuable today. We have had several inquiries and several long conversations with some of the biggest players in the industry. And Will and I and the Board made a decision to not sell the data center. Why did we make a decision? Because we think that the data center is valued on a real estate play where we think in reality, they should be monetarily infrastructure for AI. And the monetary system around infrastructure for AI is much more valuable than it being a real estate play. Selling today would be like selling power, systems power generation at the beginning of the electrification of the United States. We think there's just too much opportunity there. We also think we're building a platform. We started with Bitcoin. And it's been a rough start. Will and I will tell you that we've had to fight being debanked more than 10x to be in the Bitcoin business. We've had to fight to stay alive. It's only recently with the Trump administration that Bitcoin has been embraced in a bigger way. Being a public company in mining, Bitcoin has been incredibly difficult, and we've been basically attacked for that. We're not saying feel sorry for us, just understand that when we started Bitcoin mining such a long time ago, it's been a difficult road. We've expanded into AI compute. Those of you who follow us know we have customers inside the data center that are using NVIDIA servers. So we're in AI compute. And now we're taking the next step and integrating robotics directly in the data center. Gary, let's switch the slide. The biggest constraint in AI today is not compute in our opinion. It's not capital, it's real-world data. The industry is trying to build embodied AI systems, but there isn't enough real-world interaction data to train the models. We're talking about a gap between millions and hundreds of millions of hours. And if you don't follow that concept, it is a concept of whether the robot will be manipulated in a way to pick up a glass, to set a glass down, to pick up a bowling ball to mop. All the real-world activities that you see today that we do, whether it's torque, the way you twist your wrist, all that data has to be captured by either what they call visual data or what ego-centric data. Ego-centric data is when a human being has something connected to their head, maybe their hands are in gloves, that data is important and the importance of actual robots doing it and collecting that data visually is important. We simply asked ourselves a question, Will and I, and that is, what if the data center could just not process data, but create data. So I need you to rethink the idea of a data center with NVIDIA servers. We have NVIDIA servers. We have Bitcoin mining taking place. This is a very big property, 617,000 square feet with an additional 20 acres available we can build on with an additional 48 we have under contract. This is a big facility. Gary, please roll the video. We're going to watch a video, everyone about our thesis, sort of our manifesto on what we're doing in Michigan. The audio is not playing. The audio is not playing. There is no audio, Gary. [Presentation]

These are not just words that we're saying. If you look and do your own research about AGIBOT, they are the largest manufacturer. I know that's hard for people to believe because they think, well, Tesla has something. Yes, Tesla has robots for Tesla, where that data is not being shared with the public. They're for their own robots, and they're not at scale yet. Actually, 5,000 robots were produced last year for AGIBOT, which is the top manufacturer of robots in the world. And we are very sensitive, Will and I, as Americans and as well as a marine, we're very sensitive to the data. And so what we've negotiated with them is a supply chain opportunity to bring the robots into the United States. What we're proposing is some of the robots you've seen are actually real robots. They're not fake or AI generated. Now in that video, some of the scenes were AI, but the robots you see here are real robots. And the thesis here is that not only are we're going to collect data, but we're going to sell these robots under the Omnipresent robotics name. And you probably ask yourself, well, who would be the buyer? Think of large frontier models. Now I got to be very sensitive to who I say, but you can look them up. You know the OpenAI type of model out there. There's a lot of people competing for this space and the access to the robots is a lot harder than you think. Getting -- yes, Figure AI has, what, 15 or something like that. They're just not available. We will be ordering hundreds of them, and you'll be able to see them in real time. We'll be demoing them, we'll be using them. We'll be selling data. We'll also be selling agentic data or data that is ego-centric. So it's not a robot lab. It's a fully integrated platform that has AI compute and inference that has the ability for us to do colocation, all the typical things that a data center does, but this is such a big data center that 100,000 square feet will be set aside. So we'll be focused on better models, faster chips, more compute and high-quality real-world data. Now as we build this out and we think about the separation of Ault Capital Group, what ends up happening in Hyperscale is Hyperscale Data ends becoming a big campus. There's a lot of opportunity for this campus up in Dowagiac, Michigan. And I know I've seen a lot about the mayor, who we respect. We respect all the local authorities there. And we believe that we're going to create more than 500 jobs in that location as we build it out. And I'm going to turn it over to Will Horne about execution, kind of what's happening on the ground there. Let's go to the next page if we can, real quickly. Listen, before we go there real quickly, there is an opportunity for you to look at the total addressable market. It's a very large addressable market. We've heard Elon Musk and others in Figure AI talk about this, that there'll be more robots than there will be humans by a multiple of 10 or something or 5. There's no way to forecast how big it is. You can see Morgan Stanley says how big the market is and the lawyers would probably prefer I don't even talk about the size of the market. But we all know that if there's a basic theme here that AI, robotics and blockchain, and we're in all of them are the dominant technologies of the world here. We have -- this deck is available. We did file it this morning under sec.gov. You can find it on the SEC website. We go to hyperscale.com. And with that in mind, I'll turn it over to Will Horne.

Thanks, Todd. So what you've seen so far is grounded in infrastructure we already control. We have, as everybody, I think, knows, 617000-square-foot Michigan campus, and that is designed to support AI and robotics workloads. It's our goal to provide an energy-conscious data center while maximizing reliability and performance at the Michigan campus. We make a point of hiring consultants and contractors locally and focus on developing long-lasting relationships and partnerships that we believe produce a positive impact for clients, colleagues and the community. Currently, the Michigan campus is about 30 megawatts. We believe it has the ability to expand to about 340 megawatts. There's multi-carrier fiber connectivity there, and it's a secured controlled enterprise environment. The campus is designed as a Tier 3 data center, which means it offers availability, high availability and redundancy. Operational continuity is maintained even during maintenance, and we expect uptime to be greater than 99.9% of the time. The facility has redundant systems and any Tier 3 facility has that. And what that means is there's multiple paths for power and cooling and that allows uninterrupted service. So if one system were to fail, the system continues to work. It also has N+1 redundancy and this design ensures that there is at least one additional component such as a power supply beyond the required systems necessary to support the full IT load. So in addition to service continuity, we also try and focus on environmental compatibility. Problems related to pollution, the greenhouse effect and global warming are the main challenges. The Michigan campus is being designed with systems that were engineered to drastically reduce direct and indirect CO2 emissions in the atmosphere and limiting the carbon footprint of the data center. As far as the current development, not inclusive of the robotics initiative, we are going to begin taking a significant amount of delivery of long lead components such as chillers, uninterruptible power supplies and coolant distribution units for the campus during this quarter, which ends on June 30. These components will support approximately 10 megawatts of new critical information technology load. We selected Vertiv as our primary equipment supplier. And the reason we did was because Vertiv has one of the largest factory-trained service force with more than 3,300 field engineers. As such, they have the capacity to provide the comprehensive level of service and support that we require as a Tier 3 facility. Within the facility, we've identified approximately 100,000 square feet dedicated to robotics operations and data workflows. What makes this powerful is integration. We are building a closed-loop system. Robots are going to generate real-world data. The data is processed and validated on site. models are trained and improved and the insights feed back into deployment. That loop is what allows us to move beyond hosting and into data creation and validation. From a revenue perspective, this opens up multiple streams, teleoperated data collection, data validation, model training support, compute tied to robotics workloads, robotic sales and deployment, ongoing service and support. None of these additional multiple revenue streams have been reflected in our forecast of 2026 top line revenue approaching $200 million. We also expect to create over 500 jobs in Michigan, as Todd mentioned earlier, over the next 3 years. These future jobs, combined with our history of hiring locally and environmentally compatible design will have a positive impact on the Michigan community. Todd?

Let's take a step back. Nobody on this call that has followed us for years wouldn't say that we've had some stops and starts. And if I were to do it over again, I don't know that I would do it the same way. I don't know that I would put everything under one roof with a goal of having the global holding company. It's clear the market has rejected that. But what we've been able to do by staying alive with the help of Henry Nisser, our General Counsel; and Will Horne, who's dedicated an entire team, Ken Cragun, the CFO, James Turner, everyone that works at the company, Joe Spaziano, you guys know him as Crypto Joe, we've been able to continue to invest. And there's a really key important component here. You heard Will talk about we're getting delivery of long lead items that we've already acquired. So we think we have a very competitive advantage. Michigan has been invested in for 4 years. So as we look at the fact that we didn't want it to take this long being debanked and all the things that we've had to deal with a low stock price and difficulty raising capital at scale and getting a proper market cap, we've continued to fight to develop, and that's where you've seen a pullback in our revenue last year and now an acceleration of revenue, an acceleration of the data center, an acceleration of the lender, the Crane company doing so well. So we really believe in the assets that are held by A Capital Group that are under Hyperscale, but we really believe in the Michigan location. We believe in the people there. Jay Looney, who is the CEO of the data center, I think he's the President of the data center. The guy who runs it has more than 25 years of experience in running data centers and Brian there have spent a lot of time building out what is going to be, we think, a flagship campus. We're going to redo the campus in a big way. And one thing, Will, I want to talk to you about so that everyone can understand. If you were on this call, and I see 110 of you, if you think about an AI data center like a CoreWeave, we're talking about -- if you're going to build out a 300-megawatt campus, you're talking about billions of dollars of NVIDIA servers and power and infrastructure. We think that there's a better CapEx story for us where we can build out the power slowly. And at the same time, believe it or not, the robotic inference that's coming, the robotic data collection is actually less CapEx than it would be to build it all out at once. Will, do you want to give them any commentary on your thoughts about how we're taking it slow with our CapEx, but we've already invested so much money. But I mean, how much more could it be? It's a big number when you look at these big AI data centers being built.

Of course. So right now, as we said, we have about 30 megawatts of available power today, which means from a critical IT load, that's about 20 to 21, maybe 22 megawatts. We're splitting it up right now into 10-megawatt increments. We're not going out putting in 300 megawatts tomorrow. Really, if we build this out to the full potential of this facility, which we believe is 340 megawatts, that's really more like a 5- or 6-year plan. It's a long-term plan. And that requires billions, as Todd mentioned. But in the short term, and I'll call this the 3-year plan, we're looking at significantly less. We're looking at couple of hundred million dollars to really build out both the robotics side of it as well as the data center itself, the main components in order to support inference.

Yes, I think it's important to point out to everyone that we've been reporting between, call it, $80 million and $100 million of cash and Bitcoin on the balance sheet and restricted cash. So we're in a much different position than we used to be. A lot of our CapEx and acquisitions has already been spent in terms of where we got today. So we're starting off in terms of robotics from a much better place than we were before. We obviously have the largest partner in the world. I once again, I want to iterate all of the data we bring in, all the robots we bring in, that data will stay in the United States. It will not be sold back outside the United States. That's a pledge we've made, and we intend to keep that pledge. So we're very comfortable that we can get through any issues that pop up with this. But one thing I want to do, and I want to make sure the lawyers will hear this, that we believe we will have robotic revenue starting in less than 60 days. We believe we'll start selling robots starting in less than 60 days. And we believe there are partners that need the data that need the robotics for testing. And we, of course, are the first to ever bring these into the United States from AGIBOT, and we're very proud of that. This is something that's been in the works for a while. The team has done an amazing job. Can we flip to the next slide, Gary? So we've talked about the addressable market. We intend to go after this market globally. But sticking to the United States, this is going to be what we consider these large frontier models that we're going to go after for humanoids. You'll start to see them a lot more. You're going to see delivery of them. We'll, of course, be putting out news and information on them and videos as they develop, be announcing partners, et cetera. So this is really a milestone event for us. We think it's a pivot point for the company. And we think that compute is currency and the data is the multiplier. I would say, honestly, that for us, this is the first time I believe that we are in a position of strength where we have something that nobody else has, not in America. It doesn't mean people won't catch up, but there's such a big addressable market that a little fraction of the market share is still going to be a big deal in data center robotic sales of egocentric data and data from robots directly. So with that in mind, we appreciate you being here. We're going to answer a couple of questions. Gary, why don't you ask the first question? I will be happy to answer it, and so will Will.

First question is, is the Michigan facility making money?

I'm going to comment, and then I'll let Will comment. The Michigan facility does, I believe, over $25 million a year in revenue. Will, I don't know what the last number was that we reported, but do you want to comment on the revenue generation there?

Yes. So $25 million is in line with what we did last year materially. I don't have the exact number in front of me. I would say that the revenue that's generated primarily is Bitcoin mining. That is something that is going to slowly over time, transition to a I would say, a much more profitable and larger number when it comes to robotics and data center.

Yes. We mainly mine Bitcoin to keep it. I remind everyone, we have a Montana data center, too, where we do mine Bitcoin, and we'll continue to mine Bitcoin. Gary, do you have another question for us?

Yes. When is the Ault Capital Group spin-off going to occur?

All right. So I know the lawyers are on the call for this one. We have said that we believe that it will happen in '26 or '27. I think the latest forecast is the first or second quarter of '27. I would say this to you in its most simplistic form. The separation of businesses that have 700 employees spread around 3 countries into 2 public companies is not as easier said than done. I would equate this to driving down the road while you're laying the asphalt, meaning you're building the road as you're driving. And it's very difficult to do that. As an established company that we believe is going to more than double the revenue this year, which doesn't include a bunch of new initiatives. The upside on our revenue is dramatic because we don't include blockchain revenue in there to any degree. We don't include robotic revenue in there. We don't include AI compute, new AI compute that's going to happen. We definitely -- the lender is -- has a backlog. So I would say that it's been very difficult to sort of get a static scenario where we can go, okay, now we're ready to file, and we just weren't there yet. Will, do you want to comment on your thoughts on next year, later this year about the spin-off and the separation of Ault Capital Group and GPUS?

Sure. So the most recent estimate that we have is Q2 of '27. And that gives us time to complete the audit of these stand-alone separate entities. as well as all of the operational issues that will need to be addressed during that time period.

Yes. And keep in mind, one of the frustrations is as a public company, when you separate into another public company, you actually deduct the price of the spin-off in certain scenarios. And with the stock trading for virtually less than the cash and Bitcoin in the bank, it is very difficult to spin off an asset when -- if you think about it, if you just took the cash in the Bitcoin, did your own homework, what would you value a crane business doing virtually $50 million a year and $10 million of EBITDA? What would you value a hotel business that has 4 hotels in it and a partial ownership of a fifth? What would you value a defense business doing more than $40 million a year in 3 countries? We just think it's tremendously undervalued. It's hard to spin it off until you get to a point where you get at least a proper market cap for the companies and the market starts to accept the fact that there's going to be a separation. We are greatly appreciative of the New York Stock Exchange American and how flexible they want to be. We've been listed. The company has been listed there since, I think, 1996. And keep in mind one thing, too, the company started off as a power company in 1969. So Will and I respect the fact that the spin-off has to happen. We know we want to pure play in AI and AI robotics, and we're headed in that direction. Will, it's fair to say we're completely committed to that process, right? You and I can say unequivocally, we are committed to the separation of the companies. Right. So Gary, any other questions?

That's it at this time. Thank you.

I want to finish on a few notes. We are going to continue to Bitcoin mine. I think you'll see less of Bitcoin mining and more open market purchases. But when people ask the question of whether we're going to Bitcoin mine, keep in mind, we bought $19 million of new Bitcoin miners in the last, I think, 6 months. So we're very committed to mining Bitcoin. We're very committed to that being an anchor on the balance sheet. We want to get to over $100 million, and we want to continue to match as best we can our balance sheet and our market cap with Bitcoin. We are going to defend this. One of the most common questions I get is about a reverse split. Will and I, through the parent company are the largest shareholders. Myself, Henry and Will, and the people at Ault & Company are the largest shareholders. I believe we have a $56 million investment in Hyperscale. We do have the ability to do up to a 5-for-1 reverse. And I'll tell you what the lawyers have said. From the time the reverse was approved on April 10, we have 1 year to do that. It is our intention to avoid the reverse the best we can. If we're forced to do the reverse, it will not be more than a 5-for-1, that's all we got approved. We do not want to do it, and we do not believe we should do it. We believe the market is not properly reflecting the value of the forecasted $180 million to $200 million in revenue, and it definitely does not forecast what we're doing in robotics. These are real partnerships. We're going to try to be even more transparent than we've been before. And with that in mind, we appreciate everyone being here. 110 people showed up today. I hope you passed the message along. If you have any questions, you can e-mail us directly. It's in the presentation. The presentation is available at sec.gov. Thank you, everybody, for being here. Will, thank you and the team. Take care, everyone.

Thank you.