Data centers in space makes no sense • Hacker News Board

Re: Data centers in space makes no sense

#971

Post by nurumaik » Thu, Feb 05, 2026, 2:06 PM UTC

beloch wrote: post #115

I would not assume cooling has been worked out. Space is a vacuum. i.e. The lack-of-a-thing that makes a thermos great at keeping your drink hot. A satellite is, if nothing else, a fantastic thermos. A data center in space would necessarily rely completely on cooling by radiation, unlike a terrestrial data center that can make use of convection and conduction. You can't just pipe heat out into the atmosphere or build…

> makes a thermos great at keeping your drink hot

Thermos has 2 parts responsible for thermal isolation: vacuum and some coating that reflects heat radiation

Also in earth conditions, most of thermal radiation goes to surrounding air, heating it. This is not the case with vacuum, so thermal radiation will be even more efficient

Re: Data centers in space makes no sense

#972

Post by mike_hearn » Thu, Feb 05, 2026, 3:43 PM UTC

casey2 wrote: post #968

mike_hearn wrote: post #582

Earlier quoted context omitted.

It sounds hard but it shouldn't not make sense. 1. Solving cost of launching mass has been the entire premise of SpaceX since day one and they have the track record. 2. Ingress/egress aren't at all bottlenecks for inferencing. The bytes you get before you max out a context window are trivial, especially after compression. If you're thinking about latency, chat latencies are already quite high and there's going to be…

There is no "solving cost of launch" even if it was 10x cheaper, for 10x lighter data centers that's still hundreds of flights and billions just to get the raw materials to GSO for the performance of a single data center, with no gain. Keep in mind there has only been ~600 falcon 9 launches in total. What makes you believe SpaceX can ramp rocket flights up faster than we could just build nuclear here on earth? Where…

I agree it's hard to get things into orbit. The question is if it's harder than solving the DC capacity problems here on Earth.

In the 1950s everyone thought we were entering the atomic age, an era when electricity would be too cheap to meter. That didn't happen: making nuclear simultaneously safe and cheap turned out to be much harder than anyone anticipated. Eventually people gave up on the Age Of Atoms and started saying that solar and wind were better. Go back in time and try to sell that in the 1950s and everyone would have thought you were insane. Huge land consumption and it only works intermittently? Why would that ever be easier than building a quick reactor?

And so here we are in 2026. The track record of SpaceX making things quickly and cheaply is vastly better than the modern nuclear industry. That's not necessarily the industries fault, but the over-regulation issue is real. If datacenters have the same problem Musk can still win despite the huge starting handicap. And everyone I know with experience of datacenter construction has similar stories. 90% of it is about dealing with governments and electricity suppliers (but that's mostly the grid, which is the government again).

Re: Data centers in space makes no sense

#973

Post by DustinBrett » Thu, Feb 05, 2026, 3:55 PM UTC

beloch wrote: post #115

I would not assume cooling has been worked out. Space is a vacuum. i.e. The lack-of-a-thing that makes a thermos great at keeping your drink hot. A satellite is, if nothing else, a fantastic thermos. A data center in space would necessarily rely completely on cooling by radiation, unlike a terrestrial data center that can make use of convection and conduction. You can't just pipe heat out into the atmosphere or build…

What about this articles points?

https://research.33fg.com/analysis/debunking-the-cooling-con...

Re: Data centers in space makes no sense

#974

Post by mike_hearn » Thu, Feb 05, 2026, 3:56 PM UTC

catoc wrote: post #886

mike_hearn wrote: post #582

Earlier quoted context omitted.

It sounds hard but it shouldn't not make sense. 1. Solving cost of launching mass has been the entire premise of SpaceX since day one and they have the track record. 2. Ingress/egress aren't at all bottlenecks for inferencing. The bytes you get before you max out a context window are trivial, especially after compression. If you're thinking about latency, chat latencies are already quite high and there's going to be…

“just an engineering problem” Sounds a bit like that Dilbert where the marketing guy has sold a new invisible computer and is telling the engineers to now do their job and actually make it.

Not at all. It literally is just an engineering problem. Space radiators exist for decades. You can build one big enough for any heat load, it's purely a matter of cost - there are no unknown problems there.

You can argue that the current costs are too high and you need new physics or new inventions to bring it down to something more reasonable, yes. But the basic science of radiating heat in space is known. There are proposals for alternative designs that might work better. The question is how much can you build them for and what's the resulting cost profile. Which is an engineering question.

Re: Data centers in space makes no sense

#975

Post by weregiraffe » Thu, Feb 05, 2026, 3:57 PM UTC

kibwen wrote: post #705

weregiraffe wrote: post #664

Earlier quoted context omitted.

>Specifically: Starship makes no economic sense. Starship can replace Falcon 9 and probably be cheaper, if fully reusable, so more profitable. So at least some economic sense is there already.

No, that's not how any of this works. Try to think for a moment why we still overwhelmingly use non-jumbo jets for aviation in a world where jumbo jets exist.

Bad analogy. Reusable jumbo jet is cheaper than throwing away a small plane after every flight.

Re: Data centers in space makes no sense

#976

Post by MarceliusK » Thu, Feb 05, 2026, 5:18 PM UTC

Yet I do think there's a narrow use case where space compute might make sense (for instance, niche workloads colocated with space sensors), but that's very different from "frontier AI training"

Re: Data centers in space makes no sense

#977

Post by MarceliusK » Thu, Feb 05, 2026, 5:20 PM UTC

beloch wrote: post #115

I would not assume cooling has been worked out. Space is a vacuum. i.e. The lack-of-a-thing that makes a thermos great at keeping your drink hot. A satellite is, if nothing else, a fantastic thermos. A data center in space would necessarily rely completely on cooling by radiation, unlike a terrestrial data center that can make use of convection and conduction. You can't just pipe heat out into the atmosphere or build…

Radiative cooling scales with surface area, compute scales with volume and density. That mismatch is brutal for high-density workloads like AI training

Re: Data centers in space makes no sense

#978

Post by MarceliusK » Thu, Feb 05, 2026, 5:22 PM UTC

elamje wrote: post #353

I was talking to someone about this the other day. I was part of a team at NASA that developed a cooling system for the ISS and this whole premise makes no sense to me. 1. Getting things to space is incredibly expensive 2. Ingress/egress are almost always a major bottleneck - how is bandwidth cheaper in space? 3. Chips must be “Rad-hard” - that is do more error correcting from ionizing radiation - there were entire t…

On the engineering side, your list pretty much nails why this reads as sci-fi dressed up as a roadmap

Re: Data centers in space makes no sense

#979

Post by MarceliusK » Thu, Feb 05, 2026, 5:24 PM UTC

dathinab wrote: post #322

The really crazy thing is you don't need to know more then basic (non Hollywood) physics to know how dump this is 1. every gram you need to send to space is costly, a issue you don't have at ground level 2. cooling is a catastrophe, sure space is cold, but also a vacuum, so the cooling rate is roughly the infrared radiation rate. This means if you are not careful with the surface of a satellite it can end up being ve…

The cooling misconception is probably the biggest tell

Re: Data centers in space makes no sense

#980

Post by MarceliusK » Thu, Feb 05, 2026, 5:29 PM UTC

sinuhe69 wrote: post #400

As the focus here is solely on the US, and the comments focus too much on the impossibility of heat dissipation, I want to include some information to broaden the perspective. - In the EU, the ASCEND study conducted in 2024 by Thales Alenia Space found that data center in space could be possible by 2035. Data center in space could contribute to the EU's Net-Zero goal by 2050 [1] - heat dissipation could be greatly en…

The danger is that these nuanced, legitimate use cases get rhetorically stretched to support much bigger claims. That's where skepticism kicks back in.