Cowboy Space Wants to Put an Energy Grid in Orbit

Most people know him as the cofounder of Robinhood, but he's also the founder and CEO of a company called Cowboy Space that wants to put solar powered GPUs, into low Earth orbit and eventually use, and purpose built rocket for the upper stages to remain in space as actual data centers themselves. We just had a little bit of a test of that. You put your first hardware into orbit, Baiju, at reason one. Give me a sense here. I know it's been... Gonna be off, what, maybe a few more days or weeks before you have some communication with that, But can you give us an update where reason one is and when you expect to actually see a proof of life of concept? Yeah. Thanks. First of all, thanks so much for having me on. Lot of big stuff going on at the Cowboys Corporation. I started the company about two and a half years ago, maybe a little bit more now, with the goal of building an energy grid in orbit. And my thesis was that energy would be the next major industry to exist in outer space. And, obviously, my background is in bringing commercial activity and and capitalism to the pockets of millions of Americans via Robinhood. And so while the company started out with the idea of actually collecting power in space and beaming it down to the ground, We're actually gonna be demonstrating that for the very first time with our technology in the coming weeks here. So reason one is our first satellite went to orbit on a SpaceX rideshare. We've made contact with it. We're we're talking to it every day right now and getting ready for that first demonstration of optical power transmission from space to Earth, which... Just to put that in perspective, we're talking about kind of what to me feels a little bit like a magic trick is you collect electricity in outer space, and then you generate electricity on the ground. And you use free space optics and lasers to carry the energy from space to Earth. Well, we'll get break this up. It it is. I mean, scientifically, yes. But I do wanna ask you just from a business perspective, an economic perspective, investment perspective, is this gonna be a more economical solution to our data center build out rather than, you know, the terrestrial solution that is underway right now? Yeah. So we actually... We believe that there's a path to doing this, and that's one of the core ideas behind what we're pursuing. So reason one, first mission goes to space... Went to space, very recently. That carries some of the core technology that will sort of transform into what we're doing with orbital data centers. And our our long term goal of this is to actually build a bespoke launch vehicle, so a rocket purpose built for putting data center payloads in space. And what that means is that the rocket that takes the data center to space, given the unique applications of data centers in space, actually transforms into the satellite. Mhmm. And by doing so, we get to reuse a lot of the components on the actual rocket itself. Mhmm. For example, using the fairing of the rocket as a functional element. So when it gets to space, it unfolds and actually is a radiator that helps dissipate the heat out. So all of these innovations that lead to a uniquely different approach to putting data centers in space. Well, on the on the scientific side, though, one thing I am curious just is about, is about the energy. I mean, you know, basically, every megawatt you put out there is, to a certain extent, also gonna be a a megawatt of heat source as well. And, obviously, the dynamics on Earth are a lot different in space when you're in a vacuum. So how has that problem been solved if it if it has? That's right. That's right. Actually, dissipating heat in space is something that every spacecraft has to do. There's a thermal budget. If you've got electronics on your satellite and you're generating power, guess what? You're generating heat. So not not a novel problem. Every space satellite, has some measure of heat measurement, man have been on or or the vast majority of them do at the very least. Mhmm. Our approach to dissipating the heat that these very large power hungry chipsets will generate is actually using something called radiative heat transfer. So using large sheets of metal, heating them up, and letting them actually glow the heat away. And so you might ask yourself, like, well, what is that? Is that science fiction? Yeah. And it's actually a pretty common thing is when you see something that gets really hot and it starts to glow, that's a version of this. We're just having to use that method to Yeah. Get rid of a lot of heat. But but with regards to actual... Having an actual GPU up there, does that mean that the GPU... Can the GPU operate continuously, or is it basically gonna have to power down or or reduce power in order to cool down at certain times? The goal is to be able to operate them at steady state on a nearly continuous basis. And and the interesting thing about doing data centers in space, and and I think one of the core ideas behind this, is that at low Earth orbit, you can get a near continuous illumination from the sun of solar panels. So you have this source of energy that on the Earth is got this sort of inherent intermittency to it. Yeah. And actually by putting it in orbit, it kind of addresses that head off. How are you gonna address potentially assuming you actually get to commercial scale with commercial customers, the issue of obsolescence? I mean, you know, if I build a data center here on Earth and that GPU is outdated or the CPU is outdated next year, you know, I just send a guy in there and swap it out. What are the economics of doing that when you have something in orbit? Yeah. Our plan is to actually run these things pretty much to end of life. And so we actually have seen this with, with older chipsets today. Chip sets that came out even a handful of years ago are still actively used today, for AI workloads. And I think it's simply a byproduct of the fact that the demand is there. And, frankly, if the chips that already exists, it's better than one that hasn't been delivered yet. And so we we actually think that the demand for a compute is so high that these chipsets will continue to get used until end of life in orbit. They may have, less glamorous jobs to do, but Mhmm. I think they'll still continue to be useful for compute.