# The Space Technology That Could Change History Forever

Markdown copy of https://realalpha.app/v/UnF8-b3xRy0/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Baiju Bhatt argues that orbital data centers could meet rising computing demand without adding pressure to Earth’s power grid, water and land.

- Show: The Shawn Ryan Show
- Hosts: [Shawn Ryan](https://realalpha.app/people/shawn-ryan/)
- Guest: [Baiju Bhatt](https://realalpha.app/builders/baiju-bhatt/) is founder of Cowboy Space.
- Published: Oct 27, 2026
- Length: 59 min

## The best minutes

- [31:43](https://realalpha.app/v/UnF8-b3xRy0/?t=1903) [Rocket parts can double as data center hardware](https://realalpha.app/s/rocket-parts-can-double-as-data-center-hardware/) (7 min)
- [10:45](https://realalpha.app/v/UnF8-b3xRy0/?t=645) [Orbit could ease data centers’ energy and land demands](https://realalpha.app/s/orbit-could-ease-data-centers-energy-and-land-demands/) (8 min)
- [46:57](https://realalpha.app/v/UnF8-b3xRy0/?t=2817) [Repeatable launch sites could scale space infrastructure](https://realalpha.app/s/repeatable-launch-sites-could-scale-space-infrastructure/) (5 min)

## Also worth watching

- [28:35](https://realalpha.app/v/UnF8-b3xRy0/?t=1715) [The first engine prioritizes reliability and fast production](https://realalpha.app/s/the-first-engine-prioritizes-reliability-and-fast-production/) (3 min)
- [22:18](https://realalpha.app/v/UnF8-b3xRy0/?t=1338) [Each Big Iron is designed to carry five GPU racks](https://realalpha.app/s/each-big-iron-is-designed-to-carry-five-gpu-racks/) (3 min)

Skip the sponsor reads, the opening event setup and team roll call, and the closing launch video.

## Summary

Bhatt says he founded Cowboy Space about two and a half years before the event, initially pursuing space-based solar power. The company had put its first satellite in orbit shortly before the event and announced a shift toward orbital data centers and a rocket to deploy them. He also recalls working with Andy Yarmola and Vlad on the early team that became Robinhood.

Bhatt argues that orbital data centers could add computing capacity without using Earth’s land or local power grids, relying on near-continuous sunlight and radiative cooling. He says Cowboy Space’s Big Iron design turns a rocket’s upper stage into a satellite, combining compute, solar panels and hardware that can serve multiple purposes. Bhatt says the economics depend on building rockets and deploying them at scale; the company targets a first flight before the end of 2028. He also describes the challenge of coordinating many technologies and plans to offer third-party payload launches to expand commercial access to orbit.

### Minute by minute

- [1:25](https://realalpha.app/v/UnF8-b3xRy0/?t=85) **Cowboy Space began with orbital power.** Bhatt says space was a lifelong interest, shaped by growing up around his father’s NASA work. He founded the company to pursue space-based solar power, arguing that the space industry needed more commercial activity instead of relying mainly on government missions.
- [6:02](https://realalpha.app/v/UnF8-b3xRy0/?t=362) **The company adds its own rocket.** Bhatt says Cowboy Space built and launched its first satellite, then announced a shift toward orbital data centers and a rocket to deploy them. He argues that controlling launch economics is important if the company wants to make the system cost-competitive.
- [10:45](https://realalpha.app/v/UnF8-b3xRy0/?t=645) **Orbital solar avoids terrestrial constraints.** Bhatt says satellites in sun-synchronous polar orbits can receive sunlight nearly continuously, increasing solar output. He argues that orbital facilities would avoid competing for land and natural resources on Earth and would not draw power from local grids.
- [14:22](https://realalpha.app/v/UnF8-b3xRy0/?t=862) **The constellation targets data-center scale.** Bhatt describes a planned 20,000-satellite constellation, with each satellite producing about a megawatt of peak power and the full network about 20 gigawatts. He compares that scale with U.S. data-center power use and says spacecraft manage heat mainly through radiation; he later compares a 20.6-gigawatt orbital grid with Tennessee’s power grid.
- [19:38](https://realalpha.app/v/UnF8-b3xRy0/?t=1178) **The rocket becomes the satellite.** Bhatt calls the design Big Iron: its upper stage unfolds into a satellite with solar panels, radiator surfaces and compute hardware. He says it is larger than a football field, weighs roughly 15,000 kilograms, produces a megawatt—enough for about 800 homes—and is budgeted for five racks and six laser links.
- [26:48](https://realalpha.app/v/UnF8-b3xRy0/?t=1608) **One engine architecture targets speed.** Bhatt describes an 86-meter rocket with a 4.7-meter diameter, nine engines on its booster and one on its upper stage, using liquid oxygen and methane. He says it is sized to take 15 metric tons to orbit; its Clydesdale engine is designed for reliability and fast development, with about 300,000 pounds of thrust per engine.
- [31:43](https://realalpha.app/v/UnF8-b3xRy0/?t=1903) **Reusing hardware could improve economics.** Bhatt says orbital computing only matters if GPU-hour costs can compete with terrestrial compute, though the two systems use different infrastructure. His design reuses the fairing as a radiator and ammonia as both electric-propulsion propellant and a cooling loop. He says the booster will be reusable and the team aims to reuse other rocket parts.
- [38:19](https://realalpha.app/v/UnF8-b3xRy0/?t=2299) **First flight target is 2028.** Bhatt says the team estimates that terrestrial compute costs will rise while orbital compute gets cheaper, and he wants to beat its 2031 estimate. He targets a first flight before the end of 2028, with a full-engine hot fire around the middle of the following year, while noting that many things must go right.
- [42:06](https://realalpha.app/v/UnF8-b3xRy0/?t=2526) **The team prepares to manufacture.** Bhatt introduces team members, including a CFO with prior data-center work at Amazon and a rocket-propulsion lead. He describes a 260,000-square-foot facility in Kent, Washington, as the company’s first production shop, where the team will work through the practical demands of building the vehicle.
- [46:30](https://realalpha.app/v/UnF8-b3xRy0/?t=2790) **Launch sites could send rockets rapidly.** Bhatt says each Trailhead launch complex would have five mounts, with four active while one is refurbished, and could launch rockets two minutes apart when a window opens. He estimates about $1 billion to establish each site and says each could produce 660 megawatts of compute per year. He frames repeatable launch facilities as commercial infrastructure that should not depend on one or two companies.
- [53:50](https://realalpha.app/v/UnF8-b3xRy0/?t=3230) **Bhatt plans third-party launch services.** Bhatt says the main challenge is coordinating resources and maturing technologies as a startup spanning data centers, rockets, lasers and optics, photovoltaics, power and finance. He announces Big Iron Express, which he says can carry about 18 tons to low Earth orbit, and says Cowboy Space plans to offer third-party access to orbit.

## Transcript

The Shawn Ryan Show, hosted by Shawn Ryan, with Baiju Bhatt.

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[0:01](https://realalpha.app/v/UnF8-b3xRy0/?t=1) with Aaron Brockovich all about data centers and how bad they are for the environment, how much of our energy they're sucking up and how concerned we all need to be about them. well a couple weeks later my friend called me and said he had a solution so I flew out to San Francisco yesterday getting ready to leave here heading back to Tennessee and interviewed him all about it I think this is a real viable solution taking these data centers off the earth so enjoy the interview enjoy the unveil let me know what you guys think see you soon how's everybody doing tonight just want to say uh thank you all for coming I had an awesome day with Bayju today, looking at all the new stuff that he's developing, and I've been really looking forward to tonight.

[1:21](https://realalpha.app/v/UnF8-b3xRy0/?t=81) So it's going to be a good night. Bayju, thanks for having me. Shall we, big guy? I know we got some big news you have some big big news to go over tonight but before we get into that could you tell me a little bit about you and how you came to the conclusion to found this new company yes so first of all thank you all for coming welcome to the this first ever event that we're doing with the Cowboy Space Corporation called Big Sky Rising, and we're excited to show you guys what we've been working on. But the journey here has been about two and a half years of working on the company full-time, but as my family knows, it's been a lifelong passion of mine.

[2:18](https://realalpha.app/v/UnF8-b3xRy0/?t=138) I grew up with my parents, who are sitting right there, with my dad working as a contract research scientist at NASA through my childhood, so I kind of grew up around the aerospace space world. And I think one of the biggest observations of that from childhood was that it felt like the good stuff happened before, you know? Like it felt like all the interesting stuff happened and that a lot of what NASA was when I was a kid, it felt kind of like a museum, right? And you'd be like, oh, here's an important part in human history. And it It has never really made sense to me why it wasn't still happening.

[2:58](https://realalpha.app/v/UnF8-b3xRy0/?t=178) But the journey from there to what would become Cowboy Space was much longer. So about two and a half years ago, I started this company with a vision of building the first American energy company in space. And why do that? Well, let's actually take a look at the mission of the company, which is powering humanity from the high frontier. here. The thought process behind this was looking at the commercial applications for outer space and thinking to myself, if humanity was going to be the space-faring civilization that I think all of us want it to be, it wasn't going to get there on the path that it was. The space industry, through the majority of our lifetimes before the last 15 years or so, has been for the government, by the government at government pricing.

[3:49](https://realalpha.app/v/UnF8-b3xRy0/?t=229) Every mission is driven by the government. And what that means is as political winds change, any of these things change, the space industry is one of the first ones to go. Now, to kind of contrast that with the commercial world, right? The world I want to live in and the world I want this company to help us create is one where there's a new generation of young people that are pursuing careers in this industry so you know last time we were together i think was about a year year and a half ago and you were building satellites in space to beam solar energy into earth and now you've got a little more to talk about today yeah a lot more to talk about today. So what was it that shifted focus? What was the main thing that shifted focus into this?

[4:47](https://realalpha.app/v/UnF8-b3xRy0/?t=287) Yeah. Well, let's take a look at what we've done over the last two and a half years. It's been a quick one. Companies started out with the idea of building a power grid in space and an old NASA idea called space-based solar power. Had you ever heard of that before we talked about it no yeah um it was actually born of the last energy crisis and the thought process that i had when starting the company was let's take this old science fiction idea and see if you can make it science reality like why hasn't anybody actually tried this in earnest like i you know there's got to be a way to bring this idea to life and um a couple of weeks ago We put our first satellite in space.

[5:41](https://realalpha.app/v/UnF8-b3xRy0/?t=341) So far, it's up there. It's talking to us. We're testing a lot of the functionality on it, and we're getting pretty close to actually showing off what it can do. But that might be another one of these events that we do together, which I think would be really cool. But more to come on that later. The company started in a tiny little garage in El Granada. Not a garage, actually. It was more like an apartment place. It actually was maybe just an apartment. Like, I actually don't think it was an office space. And from there, built our first satellite, put it up in space, like I said, relatively recently.

[6:18](https://realalpha.app/v/UnF8-b3xRy0/?t=378) And it was this past spring that when we announced our Series B fundraise, We both changed the name of the company and announced this bigger plan that we had around building data centers in orbit. And this even bigger idea that we were going to build a rocket to get those data centers to space. And I think a lot of people thought that that was kind of crazy. But I don't think they're going to think that it's that crazy after this. So can I show you something? I'd love to see it. We've been cooking. oh we've been cooking saw this earlier today can you describe what we're looking at here yeah this is uh the first bit of our engine turning on so for those of you that don't know we announced um in the spring that we were going to build a rocket to put data centers in space and the reason we had this thesis was because we knew that if we wanted to actually make something cost competitive here we need to control the economics of it now people said oh beiju you're gonna build a rocket like you you and what army well this army um actually some of them are here um uh a little i don't know if you want to do this now but we have our head of rocket propulsion here, Mr. Warren Lamont.

[7:51](https://realalpha.app/v/UnF8-b3xRy0/?t=471) Lord knows that man will give a speech if I give him a chance to, so we'll save that for the end. What you're looking at here is five months of work, and it's what I would call turning hardware into data, because we blew part of the engine here. You built this from nothing to that in five months. About, yeah. Less than five months. Yeah. It's been a quick one. The goal, and we're going to talk a lot more about what this is, just so you guys know what we're looking at here, this is at NASA Stennis, and it's a testing facility for rocket engines. This is the gas generator. It's what drives one of the fundamental parts of the rocket engine called the turbo pumps.

[8:36](https://realalpha.app/v/UnF8-b3xRy0/?t=516) we'll talk a lot more about what the rockets do how they work what the different parts of them are but I think this I mean this is about as real as real gets right we're doing it we're building the engines and we think that this is going to be one of the critical pieces of technology that's going to let us get to orbit relatively quickly so why sorry why have you decided to to not only transition to data centers, which we'll get into later, but I'm curious, why are you building the rockets themselves to deliver the data centers to space? Why have you decided to do that? Great question. And I think that there's a few aspects to this, but I think fundamentally is a lack of commercial access to orbit right now, right? We know that there's one major company that controls the majority of access to space and they've said publicly that they're going to be pretty focused on their own missions to orbit. We want to build a company that builds an independent data center in space, an energy grid in space. And to do this, because the energy industry and all these industries are so commoditized, the economics really matter and controlling the economics really matter. And so this is like one of the critical parts of this

[10:01](https://realalpha.app/v/UnF8-b3xRy0/?t=601) And we want to do a lot of it. We want to build a lot of data centers in space because I think that's what the problem calls for. But rather than actually talking about it a little bit, why don't we show people what that means? Well, first of all, let's talk about why. Why on earth are we talking about building data centers in orbit like like why on earth is this even remotely a good idea or worth talking about well there's a bunch of reasons why and i i hope that by the end of the day today most people in this audience would agree with my viewpoint that it's actually a pretty good idea to do i mean a lot of i just did an interview with aaron brockovich who is a lot of of Americans, I think a lot of people in the world are really concerned with the amount of energy that the data centers are consuming.

[11:34](https://realalpha.app/v/UnF8-b3xRy0/?t=694) They're concerned about the water supply. There hasn't been a lot of talk about putting these things in space. I think, I mean, I think it's a big problem you're tackling here. Yeah. Yeah. There's a few advantages of doing this stuff in space and you hit on a couple of them. And I'll just kind of go over this quickly so people know what we're talking about. What you're looking at here on the screen, little orange lines represent where the satellites would be in orbit. And you can see the way that the satellites are positioned, they're illuminated by the sun nearly continuously. So we put our satellites in these polar orbits.

[12:15](https://realalpha.app/v/UnF8-b3xRy0/?t=735) They're called sun-synchronous orbits. Some people call them the terminator orbit, which I think is pretty cool. But in these orbits, your satellites are illuminated nearly continuously. And so what we're talking about doing here is taking what was already one of the cheapest forms of energy production, which is traditional solar, and putting it in an environment where it produces even more power per unit solar panel. So there's clearly something there that, if we can harness it in the right way is a driver of much more energy and ideally much more economically. So that's the first one. The second one is as we look at data center build outs, it's fundamentally a real estate problem. It's real estate development, it's procuring the natural resources to power the data

[13:05](https://realalpha.app/v/UnF8-b3xRy0/?t=785) **Baiju Bhatt:** centers, whether it's energy, water, land use rights, we are not using anybody's land in outer space, in case anybody, like, didn't get that part of this.

[13:22](https://realalpha.app/v/UnF8-b3xRy0/?t=802) And the last part about this is, is we, this is, I mean, again, should not be a huge surprise to anybody but there's not literal Power plugs coming down from the satellite into the power grid in your local neighborhood. This has no effect on your power grid. And so the idea here is, is that we as a country, I think, can have our cake and eat it too, right? We can both lead in frontier AI and not have, you know, these negative effects on American houses that I think is pretty universal in 2026. yeah it seems um a lot of strain is going to come off the power grid i mean i think the power grid is something anybody who's educated on is concerned about it yeah and so i mean is is one of these satellites going to be able to produce enough energy with just solar to to power these what's inside of the satellite the data center It's a great question. And we're going to talk about how we get it to the size that matters.

[14:31](https://realalpha.app/v/UnF8-b3xRy0/?t=871) Let's actually go there now. So earlier this year, we announced to the world that we're building a constellation of these things in outer space. So it's 20,000 satellites. Each one produces about a megawatt of peak power. It's a constellation in orbit that produces about 20 gigawatts of power. It's a lot of power. For the audience that doesn't know what 20 gigawatts of power is, can you give me a comparison of how much power that actually is? Yeah, I think we can do that. This is what data centers use. Wow. What am I looking at here? What is this graph? So what we're looking at here is power drawn in gigawatts by data centers per year.

[15:30](https://realalpha.app/v/UnF8-b3xRy0/?t=930) And you can see here, it's exploding, right? The amount of energy that data centers are going to use, I mean, in 28 and 29, it's, I mean, you can look at the numbers of how much power data centers are using even in 2023. three. And the effects that it's having on people's water bills, people's electric bills, I know you've had a lot of firsthand experience with this stuff and talked to people that have lived with it. It's real. And we see it in American politics. Data centers are pretty unpopular. But they don't have to be because we can build them in outer space. So this is roughly the scale of what we're contemplating with our constellation build. It's building a power grid in orbit that generates the same amount of power that all data centers in the U.S. used just three years ago. Wow.

[16:27](https://realalpha.app/v/UnF8-b3xRy0/?t=987) How are you going to cool these things? I mean, a lot of people, a lot of people are concerned about the water. The data centers pollute the water, they use all the water, worried about the depletion of our aquifers I mean it's a massive concern luckily water is not very good at cooling stuff in space and so there we'll talk a little bit more about how we keep these things cool in space the physics of it are pretty interesting it involves actually using the coldness of outer space and the temperature differential to emit heat in outer space. So just a little reminder, the way that we all on Earth experience heat transfer totally different than in outer space, right?

[17:18](https://realalpha.app/v/UnF8-b3xRy0/?t=1038) On Earth, you experience heat transfer through conduction, meaning heat transfer, through convection, fans blowing, and through radiative, like in the case. of something getting so hot that it glows in space there's no air and you're kind of not really touching very much and so that radiative heat transfer is your main method of dissipating heat from the system and it's not this like it's kind of interesting because when people talk about data centers in space this concept of how do you dissipate the heat that they generate is front and center but when you talk to people that actually work in the space industry it's like every single spacecraft has a thermal budget.

[18:04](https://realalpha.app/v/UnF8-b3xRy0/?t=1084) Like, you can't put stuff in space and not think about how it dissipates heat for this very reason. And so we can get to a little bit more of that in a little bit. Let's talk a little bit more about how big this thing is. I think you know a little bit about this state. That would be my state. Yeah. So we're contemplating building a power grid in space that's about the size of the power grid of Tennessee. Wow. 20.6 gigawatts. And when you look at numbers this big, I think it should become pretty clear that this amount of energy is just not available on Earth. Like, we would have to fundamentally rethink the way the entire electrical grid works.

[18:58](https://realalpha.app/v/UnF8-b3xRy0/?t=1138) And there's definitely people out there that are doing it, but as we get further along in this, we'll talk about why we think that not only does this solve the sort of macro of the problem, but it also makes it fundamentally more economical. So, we believe that space is inevitable. It's a natural place for collecting electricity. There's an abundance of space to do it in. the method of heat transfer is actually well-established. And then the real challenge becomes how you do it and what exactly you're doing to make this happen. So you guys want to see what our satellites look like? I would love to see what they look like.

[19:52](https://realalpha.app/v/UnF8-b3xRy0/?t=1192) So there's 20,000 of these things that are going to go into space. Each one is pretty big. 20,000? 3, 2, 1 Introducing today for the very first time, the big iron. it's a fundamentally novel approach to to computing in the or to to rocketry in the modern era in our opinion because rather than using the satellite the rocket as a cargo vehicle to taking a payload to space our approach is fundamentally different the cargo vehicle is the payload in space and We'll explain exactly why we did it this way and why it really matters. So the cargo payload is actually going to transform into the satellite that collects the power.

[21:43](https://realalpha.app/v/UnF8-b3xRy0/?t=1303) So the idea is that you have a two-stage rocket where the upper stage of the rocket itself transforms into the satellite. And this is to differentiate from other space companies that will pack the top of a rocket with a stack of satellites, imperial-class satellites, whereas what we're doing is actually using the whole vehicle as a satellite. And so it's a little hard to tell from this, so let's kind of show you a little bit about what is going into one of these things. So there's the big iron. So you can see here it gets to space. This is the whole upper stage of the rocket. We deploy out the solar panels and then the shell of the rocket, the fairing, folds out and becomes the method of heat transfer. It's pretty big. How big is it? It's bigger than a football field. It's 15,000 kilograms, roughly. It produces enough power to power about 800 homes. Like, it's a big goal thing. And when we talk about power in the data center world, I think sometimes we lose sight of just how much power we're actually talking about here. This is a megawatt of power. This is like a pretty significant amount of power. We generate it with pretty traditional solar panels.

[23:30](https://realalpha.app/v/UnF8-b3xRy0/?t=1410) **Baiju Bhatt:** Each one of our big irons, additionally, has six laser links on it to be able to transmit data to neighboring big irons or to transmit data down.

[23:43](https://realalpha.app/v/UnF8-b3xRy0/?t=1423) We're targeting very high bandwidth laser links. Our team has been working on laser and optic development for the better part of two and a half years now. And the big relevant bit here is the amount of compute that we're able to put into the satellite. So each one of these things, we're budgeting about five racks of, for example, NVIDIA NVL72 racks. per spacecraft. But it's kind of hard to tell how big this is. So maybe this will help. Wow. Yep. The American football field. Yeah, and a big iron, which is obviously why we need a big space to build the big iron but let's let's just kind of articulate what we're doing here and in kind of the simplest terms right if you if you look at what our approach is it's building a rocket that is just big enough to build to put this payload into space and this basic approach is driven by the idea that when you're talking about compute more interconnectivity is inherently better than less right and so rather than building a vehicle that could take many different satellites to space that could do AI compute you'd actually rather have one big one you'd actually rather have one big iron. Oh, God. But as we all know, cowboys Space is not the only space company in the world.

[25:35](https://realalpha.app/v/UnF8-b3xRy0/?t=1535) There's another company out there that's also building these. And when we take a look at that architecture from the Empire, you see the Imperial-class satellites. They're just way smaller than ours. Those are smaller than yours. Tiny. Each one of them packs about one rack. Ours tracks five. And so whether you're taking a starship or a star destroyer to orbit, we think Big Iron actually delivers more interconnected compute per satellite than what the Empire is working on. And if you kind of think about what this thing is, and we'll talk more about how this thing Works and what it does this is not the like the performance and of The of the whole vehicle and the way that we get it to space is Oriented around speed to market so probably talk about that next Right how are you going to get something that big into space Yeah well it's got some big specs So a little bit on the rocket architecture.

[27:02](https://realalpha.app/v/UnF8-b3xRy0/?t=1622) It's about 86 meters tall, 4.7 meter diameter. We're sort of pushing the boundaries of what's road transportable. That's kind of how we selected it. We'll talk a little bit more about our engine architecture in a minute. It's called Clydesdale. The booster stage has nine engines on it. The upper stage has a single Clydesdale engine on it. From a rocket architecture perspective, we're going to where we think the results, where kind of the end goal is. And so from that perspective, we're starting out with liquid oxygen and liquid methane as the propellant, stainless steel rocket, and its size to take, as we talked earlier, about 15 metric tons to orbit.

[27:50](https://realalpha.app/v/UnF8-b3xRy0/?t=1670) So you can kind of think of it as sort of pushing the boundaries of what would be a road transportable rocket. And as we will get further along into this, we'll talk about the logistics of bringing these things to life. But if we started out the discussion by articulating what the problem to solve is, which is that we have to put up a ton of compute and we have to power it in a somewhat novel way, If that's the why, like, this is how you do it. A little bit more on the engine. And this is the brainchild of Warren, who, again, is sitting right there. There we go. There she is, Clyde's Hill Engine.

[28:47](https://realalpha.app/v/UnF8-b3xRy0/?t=1727) That's what you saw us testing at least a part of earlier in the video. And this is a gas generator cycle engine. This is not a racehorse.

[29:04](https://realalpha.app/v/UnF8-b3xRy0/?t=1744) **Baiju Bhatt:** It's meant to be reliable and fast to build. There's definitely more performance in this engine. We can solve that problem at a later point in time.

[29:13](https://realalpha.app/v/UnF8-b3xRy0/?t=1753) The goal with this program is to bring this capability online as quickly as possible. We'll talk a little bit more about that in a moment. But each engine produces about 300,000 pounds of thrust, which is kind of difficult to even wrap your head. How close are you to this? I mean, this is what we saw. This was one of the first videos we watched. Yeah. And then being at your facility today, talking to Warren, I mean, from hiring him to getting that, was that the first test? We did some testing a little bit before that this week, but I think that's in earnest the first test. And that was today? Yeah.

[30:00](https://realalpha.app/v/UnF8-b3xRy0/?t=1800) And the next test is tomorrow morning. At 7 a.m., right? How are you moving that fast? We are very motivated. We really, really care about making this happen. And you just can't fake that, you know? Like the most interesting things go to the people in life that are the most motivated, in my experience. engines. But rather than waxing poetic, let's talk about rocket engines. Yeah, so this one is, I mean, these things are crazy how powerful they are. It's like kind of difficult to wrap your head around it. And I asked the team to kind of try to articulate some way to explain what this thing does. And I think the best they came up with is one of these little engines here has enough thrust to hurl 50 f-150s to orbit wow yeah we were talking we're having this whole conversation about you know it's like well we're we're we're brainstorming all the different stuff we can do to to test the engines test the vehicles and there was a discussion at some point by the way there's very opinionated people at this company and if you have an idea about something there will almost certainly be somebody in the corner being like terrible idea never do it But this terrible idea was doing the hopping, like building a small rocket and doing little hops

[31:21](https://realalpha.app/v/UnF8-b3xRy0/?t=1881) and staying under a certain height limit. And the point I was told was that these engines are so powerful, and you're just like, pew, straight to orbit. So it would be a different engine. But that being said, this is the engine that's the single architecture that actually works for both the upper and lower stage of the rocket. And so we'll talk a little bit more about that in a second. What I want to come back to here, and I think the most important thing about all of this, is how do you make this economical? Meaning, we could do all of these things. Sure, we could build data centers in space.

[32:01](https://realalpha.app/v/UnF8-b3xRy0/?t=1921) But if it's wildly expensive, it's a little economic utility. And so the question that I've had from the beginning of starting this company is, what is the path to cost-competitive economics? Like, how do you get the cost of producing the output service, in this case, a GPU hour, to be on cost parity with what it costs to build it terrestrially? And if you really think about it, this is fundamentally an apples to orange comparison, right? Because on the one hand, if you want to power a GPU on the ground, you need a building, you need a power plug, you need some cooling, and you need to, like, do real estate stuff, right?

[32:43](https://realalpha.app/v/UnF8-b3xRy0/?t=1963) In the case of powering that thing in orbit, you're talking about putting methane in an engine and like building a structure and like leaving it up there for a decade or five years or whatever. All this to say the physics of how you produce the same thing are very different, but that does not, that has not prevented us from asking the question and really answering what does it cost and how low can we go. And going real low means being creative. And this is hopefully going to put a little bit more color on the idea earlier of the entire vehicle, the entire rocket is the data center. Why would we do that?

[33:24](https://realalpha.app/v/UnF8-b3xRy0/?t=2004) What's the point of that? Well, let me show you. It's so that we have.

[33:31](https://realalpha.app/v/UnF8-b3xRy0/?t=2011) **Baiju Bhatt:** We have parts on our rocket that do two jobs, so that we don't need two parts.

[33:38](https://realalpha.app/v/UnF8-b3xRy0/?t=2018) And I think this is going to be pretty cool. So I want you to think about a rocket in two forms. One, on the way to space, and two, when it's already in space. So on the left side, you'll see pictures of different parts of the rocket on ascent when it's a rocket. And the second one will show you what that looks like when it's in orbit and it's a satellite. So, first things first, the outer shell of the rocket, the nose cone, the aerodynamic shell of the rocket, has a specific purpose on ascent. It's for aerodynamics, right? In our design, when we get to orbit, the actual fairing folds out, and it starts working as a radiator.

[34:26](https://realalpha.app/v/UnF8-b3xRy0/?t=2066) And so that extra metal, the metal that we would have taken up anyways as a part of the rocket itself, does a part of the satellite's job when it's there. And in turn, less stuff you have to take up. Similarly, propellant. Our design calls for a two-stage rocket that gets to a parking orbit. And then from there, it uses a different method of propulsion to get to its destination. In this case, it's electric propulsion. And the propellant that we're using is ammonia. Now, the interesting thing about ammonia is that it also works well as a coolant. And so the same working fluid that we use as a propellant on ascent, when we're in orbit, it, we can use that same tank and the residual ammonia in it as a cooling loop for the GPUs.

[35:26](https://realalpha.app/v/UnF8-b3xRy0/?t=2126) So imagine the tank of propellant that basically pushes this thing up to orbit when it's there will then take that same tank and start running it through the GPUs and extracting heat out from them and running it out to those big wings and bringing them all out to an even temperature so they can then dissipate the heat out. Genius. Double duty. Oh, this one is my favorite. So cold gas thrusters. Rockets have cold gas thrusters. And unlike the Empire, we've actually found a practical use for cold gas thrusters that does not involve kickflips. The cold gas thrusters on our upper stage help orient the vehicle.

[36:19](https://realalpha.app/v/UnF8-b3xRy0/?t=2179) When the thing is actually in orbit, we use the pressurized tanks to push the solar booms out. Didn't even need to do a kickflip, and it's actually useful. So, now, we will be building a reusable rocket. So, the booster gets reused, and we talked about this a moment ago. Because it's one engine architecture between the two stages of the rocket, when we get to an engine that is approaching end of life, we'll take that engine and we'll put it on the upper stage because that engine, when it's done with its orbit assign, actually gets deorbited along with the tank. So the last use an engine has is to take the second stage to orbit. So literally every part of this rocket will be reused. We're trying. We're trying. And the reason this matters is because the more efficient we can get with what we take to space so Here's basically the thing.

[37:32](https://realalpha.app/v/UnF8-b3xRy0/?t=2252) The rocket produces thrust, right? Every bit of thing that you're taking to space that makes money creates the equivalency between thrust and revenue, right? And so said differently, what we want to do is we want to be spending our thrust on things that generate revenue, not on other stuff. So that means cutting out redundant parts because we're spending valuable thrust to get that stuff to orbit. So, for example, if every satellite needs a radiator, well, you could pack a rocket full of individual satellites that carries its own radiator, or you could reuse a part of the actual rocket as the radiator.

[38:19](https://realalpha.app/v/UnF8-b3xRy0/?t=2299) Does that make sense? It does make sense. Are people following this? Does this make sense? All right, cool. So, I have made the point a couple of times now that none of this matters unless it's economical. And that's what we're going to do. It's going to get cheaper every time. Yeah. Again, these are estimates. But as we see the cost of terrestrial compute actually increasing, our estimates have orbital compute actually getting cheaper. By 2031, for those of you that know me, I will be pushing to make that happen a little bit sooner. But this is what our team put on the deck. So but there's a little Easter egg buried here.

[39:13](https://realalpha.app/v/UnF8-b3xRy0/?t=2353) We're going to try to launch this thing in 2028. You're going to launch this in 2028? Yes. Our boy Warren's got his gork cut out for him. Nervous laughter. Yeah, so what does that mean? If we want to do this, we've got to get to it. And I have left the team with this mantra many a time, which is that the best way to get good at building rockets is by building rockets. and so we're going to start doing that our goal is to design a system that does a couple of things number one gets to cost parity or cheaper with terrestrial economics that was the alpha and the omega of the design prompt right it's like how do we get there what is the technical path to making that happen? Secondarily, was how do we down select the technologies to ones that are mature enough that we feel confident we can bring them online in record time? And that's what we're talking about doing here. That involves getting to a hot fire of the whole engine sometime in the middle of next year, getting something close to a booster testing between now and then as well, and then before the end of 2028, our target is a first flight.

[40:54](https://realalpha.app/v/UnF8-b3xRy0/?t=2454) Now, a lot of things have to go right for that to happen, and it's not a guarantee, but this is the kind of ambitious goal that I think is like a truth serum. It's like, if you actually want to do this, then how do you do it, right? Cut out all the noise. Like, what is the bare minimum path? What is the path that actually captures the outcome that you want without all the other stuff? And I think it's ambitious, but if it wasn't ambitious, I wouldn't be an entrepreneur. So let's show you a little bit more. Transcription by CastingWords Giddy up, giddy up. So next part of this, we're going to talk about how we do this, right?

[42:23](https://realalpha.app/v/UnF8-b3xRy0/?t=2543) Start out talking about why we should do it, Big Iron is what we're doing, and this is the how. And a big part of the how is actually the who. It's these folks. So I'd actually like to welcome all of our team members that are here to actually stand up. Give them a round of applause. extra special shout out to my homie joe this company would not have happened without him Although he didn't wear a bolo hat, a bolo tie or a cowboy hat today. Lawyer's going to lawyer. Lawyer's going to lawyer. One of the best parts of doing something like this is working with a group of people that wants to believe that they can do things that other people think are just freaking crazy.

[43:37](https://realalpha.app/v/UnF8-b3xRy0/?t=2617) and watching those people find it in themselves to do it is a magical thing so who are these people so talked about joe uh spent a career in law um was most recently at scadden arps for 16 years is that right um we've got chithra who's our cfo uh she was at avarta more recently and amazon working on data centers for a long time where is chithra there she is uh warren although I think his head's big enough right now. We don't need to give him any more love. No, it really can't be big enough. We've got Reggie, Mr. Lazer's over here. We've got Ross, somewhere in the back. Tyler. Where's that Tyler at?

[44:26](https://realalpha.app/v/UnF8-b3xRy0/?t=2666) There he is. There's that cowboy hat. Tyler and David are the two folks that joined us at the beginning of this year, and we're pretty instrumental for putting this plan together. My dear, dear friend, Tom Linford. See you back there, big guy. John Sarkis. And Andrew Yarmola, who is not here. And I actually asked to put a cowboy hat on his head because he's not here. Andrew is actually one of my oldest friends and was my classmate at Stanford. Once upon a time, three of us were working on a little company together that would go on to be a slightly bigger company called Robinhood. And that original crew of three people that worked on the company before all this stuff was me, Mr. Andy Yarmola, and a guy named Vlad.

[45:22](https://realalpha.app/v/UnF8-b3xRy0/?t=2722) Wow. Yeah. And 2027 is the year of Goblin Mode. This is our facility in Kent, Washington. It doesn't look quite this nice right now. This is more future aspirational state. The rocket's not there yet. But look, we're going to, we've got 260,000 square feet there. The point of this facility is this is our first production shop. And what we're going to do here is we're going to live, breathe, and eat the pain of making one of these things on a day-to-day basis. so it's next year is a full-on live with the pain figure out all the things that you don't know that you wouldn't have known unless you just did the thing it's yeah goblin mode what's it gonna look like to get this what you're building in there in space it's gonna involve building a lot of tents in the parking lot yeah so we can we can talk a little bit more about that in a second but first we we've talked about kind of how we get to that first launch this this facility is going to be where that happens but i also want to give um a view towards how we get to that 20 000 constellation size because fundamentally like the logistics and the operations of how we do it But these are still a couple of years out from us actioning them.

[46:55](https://realalpha.app/v/UnF8-b3xRy0/?t=2815) But we're at the point now where we're starting to design them and think through what that would be. And so we wanted to share some of that with you guys. So massed orbit. You can kind of summarize what our approach to building rocketry here is. is that rather than building one gigantic rocket,

[47:18](https://realalpha.app/v/UnF8-b3xRy0/?t=2838) **Baiju Bhatt:** we actually think it's more economical to build a slightly smaller rocket and make more of them and launch those.

[47:25](https://realalpha.app/v/UnF8-b3xRy0/?t=2845) And so our design for our launch facility, which we call the trailhead, is rapid deployment of multiple rockets at a time. One of the things that we talked about in the beginning was the fact that we're putting our satellites in this, like, terminator orbit, right? Well, the thing about it is it's the same orbit every time. And so you're launching rockets in the same trajectory, meaning that every time you're putting a data center satellite in space, your booster's landing basically the same spot, right? And so you design the launch operations around this. So rather than having a single launch mount per launch facility, we contemplate having four that are active at a time. And you launch rockets two minutes apart when you see a clear window opening to get stuff to space. And as we said before, when we get the space, the booster comes right back and we get to reuse it. The goal with this approach is speed and economics. Those are kind of the two through lines of this. And what we hope to achieve with each trailhead is producing about 660 megawatts of compute per year per site. Initial cost is each one of these sites can be stood up relatively inexpensively. We think we can stand up the production site and the

[49:17](https://realalpha.app/v/UnF8-b3xRy0/?t=2957) launch pads for about a billion dollars per facility we'll have five pads or sorry five launch mounts on the launch complex four will be active at a time while we refurbish one of them And the idea is rinse and repeat. Like, you can produce these things all around, at least the coastal US, other spots abroad as well. And I think there's one thing that's very important here, kind of in the big scheme of things, things, which is where we started this conversation out, which is the idea of commercialization of outer space. I don't want to live in a world where it's entirely driven by government activities.

[50:12](https://realalpha.app/v/UnF8-b3xRy0/?t=3012) We also don't want to live in a world where it's driven entirely by one or two companies. We need access to orbit and we need access to orbit for broad commercial activity. the bold ambitious vision for the world we want to live in is that humanity has meaningful applications for outer space and it's not tourism it's not just looking down at the earth but that there's reasons that people are building this infrastructure in the most american of senses in pursuit of their own self-betterment for the fruits of their own labor we just don't have that in space right now like you can't do that and make a living in the way that we we see the opportunity and again this is kind of like zooming out a little bit and talking about what we want to actually have happen here but even in the context of uh developing the trailheads right i think this is an interesting point that i wanted to just articulate towards the end of this which is that if you think about the future of the space industry as being one that's heavily commercial, you start looking at launch complexes a lot more like you look at airports.

[51:28](https://realalpha.app/v/UnF8-b3xRy0/?t=3088) And there's an entire industry there that I think is going to exist in years to come, free markets and capitalism of financial services providing the capital to build, you know, these pieces of American infrastructure, launch complexes, right? And when we look out into the future, space should be a place about economic opportunity, not for just the government, not for just one company, but for humanity as a whole. And to do that, we need more people going to space, and we need more people that have figured out how to make money in space. And at the end of the day, that's what we care about. So let's wrap this up.

[53:04](https://realalpha.app/v/UnF8-b3xRy0/?t=3184) 3, 2, 1 Oh, my God. Transcribed by ESO, translated by — i'm curious you know i mean like i said we met for the first time about a year and a half ago and it was power plants in space now it's data spinners in space plus delivering the payload what do you find in is the most challenging part right now um is it satellite communication is it the rockets the hardest part about this is is building something that works at scale figuring out how you marshal the resources how you stair step the deployment of resources because we're you know we're still a startup which means that we have to be very judicious about what the right proof points are, right?

[54:53](https://realalpha.app/v/UnF8-b3xRy0/?t=3293) How do we get to mature, how do we mature the right technologies at the right time and that actually show the world the progress that's behind what we're doing? I don't know if that's a great answer to it, but it's just the amount of complexity and different moving parts of the business, the different timelines that pieces of technology you're finished on, and kind of keeping a variety of different things moving simultaneously. I mean, we are a data center company. We're a rocket company. We're a lasers and optics company. We're a photovoltaics company. And with the amount of capital that we're going to end up deploying doing this, we're like a financial company as well.

[55:32](https://realalpha.app/v/UnF8-b3xRy0/?t=3332) So it... And a power company. And also a power company. And it sounds complicated when you say it like that. Yeah. But the way that I think about it is actually, it's just going to take a long time to put all the pieces together. Like, you can have a vision for something, and if you can kind of imagine how you put it all together in a year, it's not that big of an idea. like usually the big idea is you're like holy shit how do any of these pieces ever fit together and it's kind of the magic in in doing that but there's a big but i told you guys i'd have a surprise for you and there is one Introducing the Big Iron Express, we're going to fly a third party payload to space.

[56:56](https://realalpha.app/v/UnF8-b3xRy0/?t=3416) It's a beautiful rocket. It's a big rocket. It can take big and beautiful payloads to space. It's about 18 tons to low-Earth orbit, and we encourage anybody interested in this capability to go to CowboySpace.com and sign up. We're going to be offering access to orbit at a time when it's sorely missing. And we think it's actually critical to the success of our business. It's also critical to the core thesis. Behind this, which is commercial and economic abundance in outer space. That doesn't happen if one company does this. So you're a logistics company as well? Yeah, probably. Yeah. And that's all she wrote. Thank you so much.

[58:15](https://realalpha.app/v/UnF8-b3xRy0/?t=3495) That was pretty good on time, too. Yeah. Yeah. You