Applied Digital (NASDAQ: APLD) | CEO Wes Cummins | Marathon Money

Thanks for having me again, Kenny. I appreciate it. It's going great. Just busy as we can possibly be. You know, demand is great for us, which is always a plus. And just break, we broke ground on a new site in the southern part of the US, which is new for us. And so we've done everything in North Dakota. We're working on several more campuses as well. So we're just continuing to expand, you know, working on customer contracts, building out the sites, turning them on. That last part is really the key, I think, for 2026, is just making sure that we meet the deadlines we're supposed to meet for our customers from a building and turn-on perspective. We're feeling really good about that right now.

Yeah, so they are in some aspects. So you know, one of the things that we have announced and we've talked about this really for the last year. So you know, the competitive piece is we have a cloud business that owns GPUs, you know, sells GPU capacity. We announced, I think it was last April, that we were seeking alternatives for that. We've now announced at the end of December that we're going to spin that into its own separate company, and that really creates that separation, you know, to over time remove that part of our business that's competitive for the Applied Digital data center part of the business that looks competitive to some of our potential customers or our customers. So we have a really clear path to spin that, let it grow on its own. That business, you know, I think has a lot of potential as well, and there's a lot of growth out there in the cloud angle side of the business where you actually own the hardware. So excited about that part of the business and expect that to grow rapidly as well. Spinning it out creates that separation from Applied Digital, even though we'll be the majority owner over time. You know, through various methods we will not be, but it will create a lot of value, I think, for our shareholders. But it allows that business to stand alone, have its own capital stack, have it be able to raise capital on its own to go out and deploy a lot more equipment and get new customers as well, because both of those markets are very robust. And the way we go after those markets with the Applied Digital business and now what we call Chronoscale, the Chronoscale business, it's a different approach, typically different, can be different end customers. And then from an investor perspective, if you want to own data centers specifically, you can own Applied Digital, and if you want to own, you know, cloud compute, neocloud-style business, you can own Chronoscale. And it separates really those two, and they both have their own sets of positives and negatives, you know, growth dynamics and risk factors that come along with it. But I think it's a great move for us and a great way to create additional shareholder value by spinning that into its own company.

Yeah. So Chronoscale is what we previously called Applied Digital Cloud. It, you know, it looks like a smaller version of Nebius or of CoreWeave or of Lambda. And then if you look at like CoreWeave, that's in the market, they have that business, but they're also vertically integrated. That's a model they're going after, which is vertically integrated data center and compute. So when you look back to Applied Digital, you have the data center, which is Applied Digital, and then you have a neocloud business, like I said, that is more of a Nebius or a CoreWeave or that type of a business that goes out and gets customers that want actual GPU capacity. Could Chronoscale be a customer of Applied Digital in the future as a tenant for one of our data centers? Sure. But right now, Applied Digital is really focused on, you know, what we refer to as six hyperscalers for what we're looking for to grow our customer list. So that six, as a reminder, is Microsoft, Google, Meta, Amazon, Oracle, and CoreWeave. And so we have two of those six. And the goal at Applied Digital is just continue to add to that list, you know, over the next year and next two years. And hopefully fill that list out to all six of those customers over time.

Sure. So let me explain what these are. So Base Electron is a newly formed company. Applied Digital owns 10% of Base Electron. Base Electron is what we call an independent power producer, an IPP. And what Base Electron is doing is going to build power generation, initially in the state of North Dakota. And this, Kenny, it's really just forward-looking. So what we're doing right now is utility power, and we'll continue to do utility power. But what I see from a demand perspective is, you know, what I think is so large into the future that when I look out to, you know, 2028, 2029, and into the 2030s, we need to make sure more electrons get added to the grid, more power gets generated, so that we can continue to expand our campuses in the region. So in North Dakota, could go into other places in the country. And this is a way for us to control more of that power generation buildout, more aggressive in how we want to build that without the risk going to Applied Digital shareholders. So we didn't do that inside of Applied Digital. So if you look across kind of the portfolio, you have data center that's Applied Digital. What you should always expect from us at Applied Digital is building data center campuses, long-term leases with high-quality tenants. The leases we've signed so far, 15 years with extensions that can be added on to it. So you should expect more of those style of leases from Applied Digital, and that's what you're going to always get from Applied Digital: more leases, more data center campuses, and that's how we're going to continue to grow. And then if you look at Chronoscale, that's the cloud business, and you can sign what we call long-term contracts there, but they're not nearly as long as we get on the data center business. So long-term contract on a cloud business, three to five years. So that, you know, revenue could grow really quickly. The risks you deal with are more around technology obsolescence of GPUs or TPUs or whatever the cloud business owns and sells as its service. And then the third component here now is power generation. So Base Electron will develop power generation in the region. The goal will be it will work with the utilities, put power generation on the grid. It will feed Applied Digital campuses but be put in the most optimized places on the grid to feed both our campuses but also benefit the grid at large. And what we see from this is just the need for more power generation, you know, in the United States. That we're specifically focused on areas where we're building data centers, and then Applied Digital shareholders get to participate in that through ownership of Base Electron, but don't take the risk associated with building power generation, owning those assets, and, you know, the longevity and the operation. And they get some of the upside participation through ownership. But then Base Electron gets, once again, to go and have a different capital stack, different investors that put money into that company to grow it up without dilution to Applied Digital. Again, Applied Digital just participates through its 10% ownership. But grow that up and then really allow Applied Digital to continue to expand beyond the utility power that we have. You know, we're going to build utility power in North Dakota this year, next year, 2028. Then we add more power into the grid so that we can continue to expand those campuses beyond what the generation capacity is in the state right now. So it's really, Kenny, when you look at Base Electron, it's our way of controlling the power generation buildout for the future, making sure that that happens so that we can continue to expand our data center campuses. But it's looking out into, like I said, 28, 29, and beyond when we're going to need that additional power generation to come online. Right now, we have enough grid power that exists already to continue to build, you know, for the next three years. But if you're going to do power generation, you have to start planning well in advance of that and have those assets come online when we need it in 29 and 30.

That's the case. Yeah, it's starting the process. We started this process six months ago. Takes a lot of planning. The company was formed officially back in December. But the company's working through, you know, permitting, land, really sites for generation build, partnerships with utilities, all of the things that we'll need. We have a purchase that we've made with Babcock and Wilcox, which is a steam turbine on that gas that's going to deliver much earlier than traditional nat gas turbines, which is exciting. And then when you look at that evolution over time, it's not the most efficient solution. So when you're looking at kind of the puzzle pieces of what you can do from a power generation, if you're looking at that gas right now, you could do Bloom. Bloom's a great solution near-term. The primary issue with Bloom is twofold. It's very expensive versus traditional nat gas turbines. It's about double the price. And then we don't really know the longevity when it's used as primary power generation. But a great solution for many, many other aspects. And then you look at traditional gas turbines, the go-to. The issue with traditional gas turbines is if you get in line now, you're getting delivery in 2032 or 2033. So it's a pretty long lead time to get the traditional gas turbines. And then what we have is a gas boiler. So a steam turbine, you know, really probably the most proven method for generating electricity in the world. You know, most electricity still generated by boiling water, which is what source do you boil water from. And so this is in this instance nat gas, and we can get delivery in 28. It's at a great price, so about less than half of kind of what it would cost to do Bloom. But we get that delivery timeline. We have a solution that we know is going to last 30-plus years. And then if you take the next step beyond that, the most efficient way to generate electricity with nat gas is what you could call a combined cycle gas turbine. And so there's simple cycle and then combined cycle on gas turbines. So simple cycle, you're just running a gas turbine. Combined cycle, you do a gas turbine and then you pair it with a steam turbine. And so you take the heat off the gas turbine, run into the steam turbine, and that's the most efficient way to generate electricity off of nat gas. So we have the steam portion of that that'll come early, and then later we'll be able to add in the gas turbine to create additional power generation plus create super high efficiency on the locations in which we build. So it's a way of going, okay, this is a very proven generation methodology, and we can get the steam turbines now, and then we can get the gas turbines later and make it, you know, super efficient over time. But it all feeds into how can we scale our campuses in these regions beyond 2028. That's really what the focus is.

Yeah. So when you look at power generation over the near term, let's even call it medium term, eventually we have to get to nuclear power generation. It's just not going to come in the time frame that we need it and kind of the AI landscape at large needs it. So the solution for that right now is going to be nat gas. We happen to be, you know, most of our sites located in North Dakota. And North Dakota to me is, you know, it's a cold version of Texas. We have the Bakken field. There's a lot of trapped natural gas in the Bakken field that they're flaring. So it's getting burned anyways. And we're able to take that because we're working with the utility there that's building a new pipeline from the Bakken to the east. So it goes all the way to Fargo. And we'll have points along that line in which we deploy power generation. So it's burning gas in a way that generates electricity with gas that's being burned to do, you know, for the most part, nothing at this point. So when you look at the next five to seven years, it has to be a lot of natural gas generation. Some renewables will be sprinkled in. We'll do some of that as well. You need to do battery banks with that to make it, to get the efficiency, to get the load over 24 hours or get the power over 24 hours versus just when the sun shines or when the wind blows. So there's going to be a mix of what happens. And then on the way to doing nuclear, and we're not going to start talking about, you know, what Base Electron or Applied Digital will do in nuclear. We've been working on that for the past year, but it's a little premature to talk about that. But nat gas is going to be the majority of what the industry leans on for this midterm power. There's just, in my opinion, no other solution. And we have loads of excess nat gas, especially in North Dakota, that's being flared every single day.