
Orbital Data Centers and Intuitive Machines
I have always dismissed these talks about orbital data centers as just hubris, something Elon Musk and other Silicon Valley people came up with to inflate their companies values and stocks. But hearing it directly from Steve Altemus, a former NASA Johnson Space Center Associate Director, and JSC former Director of Engineering should be taken a bit more seriously. Then I saw Elon Musk's tweet today predicting ODCs will be in operation by 2029, I will take that with a grain of salt but let's say by early 2030s, most hyperscalers will have some sort of presence in space.
https://x.com/elonmusk/status/2088306926442430578?s=20
Now let's take a close look at what Steve Altemus actually said this year:
>March 2026 Call: "I think there is a lot of difference of opinion on where the actual customer base will be for on-orbit data centers and what the architecture for on-orbit data centers will be. We are studying that very carefully right now. I think what Lantaris brings to the table is this Power and Propulsion Element — the most powerful power-generating spacecraft ever built — that has the ability to be a node in a data center. If you think about data centers in particular, there is the storage element, the transmission element, and the edge computing element or the high-speed computing."
>"I think edge computing in space and doing decision-making in space is the key to the future of data centers, as opposed to replacing terrestrial-based data centers. I am skeptical about large, extremely large proliferated constellations in low Earth orbit. They have their challenges, both in power generation and in thermal management. Thinking about it with a set of large and small nodes together, maybe up in the GEO belt, is probably a better architecture, and that is where we are aiming at this point."
>May 2026 Call: "Intuitive Machines as a whole, as a combined company, has some incredible capabilities that we're thinking through as an offering for orbital data centers. The build portion of our company, the production, certainly the power propulsive element, the most highest power-generating spacecraft ever built, can be reproduced at 60,000 W. We're already thinking about, you know, 100 kW in terms of power generation. Our capabilities as a company, thinking about thermal management and managing the heat load of edge computing and the need for heat rejection in high speed computing orbital data centers is critical. As a space company, we have that skill to manage that."
>August 2026 Call: "Beyond traditional communication satellites, we believe the same high-power spacecraft architecture positions us well for the emerging commercial orbital data center market. We are currently discussing strategic partnerships to bring our expertise and satellite production capability to this burgeoning market."
Remember, the Power and Propulsion Element (PPE) that Lanteris built on the IM-1300 architecture is now being repurposed to power SR-1 Freedom mission to Mars. While other companies may talk about building these giant structures, IM/Lanteris did that already and have the expertise and processes to build these behemoth space structures.
So as companies and especially the hyperscalers start to look at scaling (see that same Musk tweet above, it was in response to a study that all AI compute scaling in 2030s will happen in space), they need to partner with experienced space infrastructure operators, and there are literally 11 companies out there with such capabilities--SpaceX and the big legacy defense contractors and then few others like Maxar/Lanteris' with the technology and flight-proven reputation. Maxar/Lanteris has close to 100 IM-1300 in orbit as I type this. BTW, Rocket Lab was just awarded its first 3 GEO satellites.
Reference: https://www.fortunebusinessinsights.com/blog/top-10-satellite-bus-companies-11136
Jensen Huang kind of echoed Steve Altemus when he said he doesn't see these proliferated orbital data centers yet, but building these big data center nodes in GEO to process space-based data (military or otherwise) and then beam relevant data optically to LEO constellations who would then beam it to Earth makes the most engineering sense at the moment, or at least in the near future. The analysts have started to pick up on that theme considering the questions about ODC's in the last few calls. If you're shooting to have something up in space by 2030, then you're already talking to partners and suppliers today and that's what he was probably referring to yesterday:
>"There's some other strategic partnerships we're looking at in terms of orbital data centers that we bid, and we'll wait to hear whether or not our experience in high-power satellites wins the day, and we win those development efforts for on-orbit data centers."
Maybe I won't take Orbital Data Centers mentions as an afterthought any longer. This is a very underappreciated part of a IM story that's dominated by the NASA Moon Base and Golden Dome contracts. It may take several months to see what role IM can truly play, but it's worth keeping an eye on it going forward. Sorry for the long post, wanted to keep it as a placeholder. I welcome your feedback.