Let's shift to data centers in space. Aaron says, 'Andrew and Ben, I've been wanting to ask this for some time.' And finally, it seems like it's hitting the news cycle just enough to catch Stratechery's attention. How do you feel about data centers in space? I've read plenty of critiques from a servicing and thermodynamic standpoint. The bull case is put nicely in this tweet. And here he embeds a tweet from Andrew Coat who says, 'Everyone is working within the same constraints. There's just one internet, just one GPU provider, same competition over data centers, energy suppliers, etc. Space changes all this. You don't need permitting for land use. There's no shortage of energy via solar. You can mass-produce satellites and they all talk to each other in a swarm via laser links.' So very futuristic stuff there. What do you think? Obviously there's the SpaceX idea.
How much time do we have? The fundamental problem, there's a lot of fundamental problems actually, there's many fundamental problems with the idea of data centers in space. Here's one: GPUs break all the time. Who's going to replace them? So you have to have a completely new architecture. You have to be massively fault tolerant. Assume stuff is just going to be dying all the time. The big issue is cooling. Space is cold because it's a vacuum. There's no air. You need air on Earth to move the heat away from the GPU or you need something else like water cooling.
It's radiation. You have to convert it to radiation. The heat is coming in from the sun, the solar panels are being turned to electricity and it has to go somewhere. It has to be a closed loop. That's how thermodynamics works. You don't get to just magically... the whole point of 'oh we have huge solar panels close to the sun'... okay, all that you're doing is for energy. Where is the energy going to go? It has to go somewhere. And so that's sort of a huge challenge to accomplish all this. You're going to need these huge panels. You're going to need this radiation sort of backside. Get it positioned in space. So the panels are one way, the radiation is the other. And that's going to weigh... it's going to be expensive. You have to get it up into space and do it there. And also, did I mention none of this is serviceable or maintainable? Like it's up there. It's up there.
Thompson, B. (2025, December 30). Could SpaceX Put Data Centers in Space? | Sharp Tech with Ben Thompson [Interview transcript]. Sharp Tech Podcast. CEOInterviews.AI. https://ceointerviews.ai/interview/924487/
MLA
Ben Thompson. "Could SpaceX Put Data Centers in Space? | Sharp Tech with Ben Thompson." Sharp Tech Podcast, 30 Dec. 2025. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/924487/.
BibTeX
@misc{thompson2025_924487,
author = {Ben Thompson},
title = {Could SpaceX Put Data Centers in Space? | Sharp Tech with Ben Thompson},
howpublished = {Interview transcript, Sharp Tech Podcast. CEOInterviews.AI},
year = {2025},
month = {dec},
url = {https://ceointerviews.ai/interview/924487/},
note = {Speaker-attributed transcript with timestamps}
}
