Chase Lochmiller, Crusoe | RAISE Summit 2026

Pumped to be here.

You know, things are going great. I mean, you know, there's a lot of demand for building large-scale AI factories and you know, I think what's incredible at this moment in time is, you know, for the first moment in history, we as humanity really have this ability to actually manufacture intelligence. That's sort of the whole premise of an AI factory is a large-scale infrastructure project that can take in electrons and can output intelligence in the form of tokens, in the form of trained and fine-tuned models that people use to produce value, people use to produce labor. And because of that, there's just been incredible demand across the board and you know, it's honestly it's difficult to keep up with on a daily basis.

Sure. I mean, look, I think there are aspects to that. Like, you know, you need a building, you need power. But, you know, the complexity really comes from the fact that what we're building are data center-scale computers. Which means all of the chips are fully interconnected on the same high-performance back-end network. You know, we use these large-scale cooling processes to cool these very, very high-density racks that are in place. And, you know, if you look at a rack in a data center from 20 years ago, that rack was probably consuming two to four kilowatts. If you look at the current Grace Blackweiler architecture, that's 130, 140 kilowatts per individual rack. And if you look at the road map, it's pretty quickly going to get to a megawatt per rack. So, you're looking at sort of a 2 to 500x increase in overall power density. That creates a lot of different mechanical, electrical, and cooling processes that need to be designed from first principles at the scale of the data center. And then, when we think about interconnecting all of these chips together, all of the chips can work cohesively together on one single computer workload. One single training process is running across the entire campus in a synchronized, orchestrated fashion that just requires an incredible amount of design and orchestration.

Yeah, I mean I think the scope, the speed and the scale are the two things that are really just mind-blowing right now. You know, if you look at the history of the internet and you know, most web applications were run out of this northern Virginia corridor where data centers have been built out for many decades. If you look at like the end of 2024 sort of in the early innings of this AI boom, the total capacity available there was a little over four gigawatts, four and a half, five-ish gigawatts. Crusoe recently announced that we've actually contracted five gigawatts of total capacity. We have over 40 gigawatts in our pipeline under development. So, you know, to put that in scale it's like we're a single company that's building the scale of all of northern Virginia, all of the infrastructure capacity today. And we're trying to do it in record-setting speed. Something that took decades, you know, we're trying to do it in a number of years. And so you know, that's really driving a lot of constraints across supply chain, a lot of the electrical equipment, a lot of the power generation equipment, across labor, the amount of electricians and plumbers and construction workers that are focused on bringing these large-scale AI factories to life is creating a lot of bottlenecks in the system. You know, if we look at Abilene, Texas where we're building a campus the first phase is 1.2 gigawatts, a 900 megawatt expansion. You know, we have over 8,000 workers there on a daily basis. So you know, Abilene's a town of 120,000 people. You know, we have more than 8,000 people just working on that campus. So, it's a really crazy moment from where it's pulling on the system in a lot of different ways and you know, trying to do it fast to stand up AI infrastructure at the demand at the speed that the demand's coming in from.

It's yes.

You know, Crusoe's been focused on this vertically integrated AI infrastructure platform. And really what that means is being able to deliver from electrons to tokens a handful of different solutions to different players across the AI ecosystem. So, you know, at the data center layer we're working very closely with a lot of big technology customers to build out those 5 gigawatts. You know, that's something that's driving a lot of focus and attention across the company. We also run a fully managed AI cloud platform, Crusoe Cloud, that delivers high performance managed clusters of GPUs to a lot of the innovators across the ecosystem. You know, this is the AI labs. This is the new startups that are building new applications and scaling those applications and needing the infrastructure to make them work.

We're building out, we're turning on, and then we're delivering a core suite of high performance services, right? So, you know, once you hit the power button, these things don't just work automatically and there's a lot of software that goes into orchestrating these large compute clusters to work cohesively and deliver results for our customers. And you know, we're really excited by, we launched this managed AI services business which is more of a developer platform to enable developers to build and scale AI applications. So, our managed inference platform enables folks to host both open source, closed source, whatever they want to get a model to us. We can host those models and instead of having to worry about the underlying infrastructure, they just get an API and they feed tokens in and they get tokens out.

You know, I would say far and away the largest source of demand is in the US for us. And, you know, folks want infrastructure based in the US. So, that is definitely the biggest focus for us just given where customer demand is coming from. But, you know, we are focused on international growth and expansion. You know, we built a big set of clusters in Iceland where we've been able to access low-cost, abundant geothermal and hydropower in Norway. We announced recently we're building a lot of new infrastructure in Israel where we have a core team of developers and we're seeing a lot of demand from the AI innovation ecosystem happening in Tel Aviv. It's just rich with a lot of really talented people.