upvote
The ocean is worse than space from every perspective but cooling, radiation shielding, and cost/ease of installation. But that just highlights how bad of an idea space-based data centers are at this time.
reply
> from every perspective but cooling, radiation shielding, and cost/ease of installation.

Oh is that all? Those are major data center concerns.

Don't forget the biggest one: an ocean-based system could be pulled up and serviced without the need for a human-rated rocket. Oh, and bandwidth/latency.

The ONLY benefit of space is that it doesn't require siting a major construction in a town full of angry residents, and it has abundant solar power. But given how much it costs to get the solar panels in orbit, that power sure ain't free.

reply
Cooling is relatively easy, you just need radiators which are passive, and essentially reduce to a launch cost penalty. You are right that they can't be serviced, but that is missing the point of orbital data centers. The whole point is that you can build hundreds of thousands of these in a factory and launch them in a scalable manner. The power, cooling, etc. comes for "free". In the long run, as the cost of the chip, launches, etc. goes down, orbital data centers will scale better terrestrial ones.

As a side note, I don't understand why I keep seeing these wrong arguments on HN repeatedly. Like everything mentioned in this thread can easily be fact checked. Radiative cooling is solved, launch costs are going down, so power costs will pay themselves back very quickly, etc.

You can argue about specifics, like chips will get more sophisticated + power efficient and fabrication will be the true longterm bottleneck, or SMRs/fusion could reduce energy bottlenecks, but talking about cooling as if convective cooling is the only option is just nonsensical.

reply
> Radiative cooling is solved,

By that logic, climate change is also solved, just built a giant radiator.

reply
"you just need radiators which are passive

"Just" is doing a lot of work there. SpaceX is planning to launch 100 GW of compute annually, that comes with ~ 2.5 square kilometers of radiator (assuming an optimistic 800K radiator temp and emissivity of 0.9, double sided)

Go for advanced carbon composites, you can do that with just 5,000 metrics tons or so of material. That's 34 starship launches just for the radiators. We haven't solved assembly, we haven't brought up power panels or core compute. Planned launch cadence that SpaceX hopes to reach end of this year: 12/year.

reply
radiators about the same size and weight as solar panels will do the trick

there is already a h100 in orbit

1GW of compute is a lot in 2026. comparing 100GW of annual compute to SpaceX 2026 goals does not make sense

if Starship launch cost predictions are accurate, data centers in space will happen within 10 years

reply
> The numbers are brutal. Starcloud’s own white paper estimates that a two-sided radiator held around 20°C would emit only about 633 watts per square meter, over 1,000 times slower than water cooling of AI chips on Earth. So, a puny 1-megawatt orbital data centre, 1,000 times smaller than the gigawatt scale of hyperscale data centres on Earth, would need about 1,600 square meters of radiator, an area roughly the size of a hockey rink.

1GW needs a pretty big area for radiation.

And in space your data centers is hard to defend against foreign actors

reply
> Radiative cooling is solved.

This is emphatically not true at any scale in which this scheme makes sense. Be careful with including too many Musk boosters in your information diet.

reply
Repairability and depreciation are the main problems. A earth data center can be repaired, depreciated, and recycled at EoL recovering some of the costs. SpaceX datacenters are a total write off from the moment they are launched.
reply
That's actually not a concern I'd have, because hardware that has been sufficiently tested and burned in tends not to fail for a very long time.

I've done builds that ran for 5+ years with virtually no physical attention, just continual degradation as hardware is taken out of service. There's also not much money to recover from 5+ year-old hardware.

I used to run AI inference GPU servers in road vehicles, which is probably an even harsher environment than a single rocket launch, and the vibration problems are real but solvable.

reply
GPUs depreciate super fast. It might last 5-7 years but it's already outdated at 2-3

Also space has more radiation

reply
5 years is a Starlink's typical lifetime. Data center satellite lifetimes will probably be shorter. Demise sooner, replace more often. GPUs get more energy efficient every year and leaving the slower, hungrier chips up there much longer than 3 years seems wasteful given the cheap cost of launch.

I think this could be done at an interesting scale even on Falcon 9 alone. If Starship does even 20% of its early design goals, it'll beat Falcon 9 and we could see orbital servers being demised and replaced every 3 years, maybe even 2, for ones with abnormally high failure rates.

Now, whether or not this will all make money in the end has a lot to do with what's going on down here on terra firma and how long it takes to get useful capacity into orbit.

(It's taken 7 years to get Starlink capacity enough for serving 10M customers. Verizon FiOS did 10M in 5 years. AT&T Fiber took 4-5 years to deploy to 10M. So, space isn't a lot slower than terrestrial.)

reply
Starlink rushed when they put the satellites in orbit, AT&T did the opposite. They did the bare minimum. So terrestrial was faster without really trying
reply
> space isn't a lot slower than terrestrial

But it depreciates faster. That fiber run is lasting for 50 years, not 5. You need 10x the installation capacity just to keep up.

reply
uhh no I dont think the road vehicle is harsher than a rocket launch
reply
The Microsoft design of filling an airtight submersible structure with argon and dropping it to the bottom of the ocean floor is the alternative design - you’re not looking to do repairs but amortize the low cost of failures across the value you extract.

The biggest issue with space is not repairability but heat - when you’re in a vacuum the only way to disperse heat is through black body radiation and that’s horribly slow compared with normal mechanisms. It means you need giant physical structures whose sole job is to accept heat from the processing core and radiate it away and have so much more material that you can radiate it at the speed you generate. It’s a huge unsolved physics problem which is why everyone is skeptical.

reply
It’s not an unsolved physics problem. Every satellite in space has to deal with it and even the ISS deals with it by having massive radiator arrays that face perpendicular to the sun.

The problem with data centers in space is one of materials science and engineering: how to make radiators large enough and effective enough to cool it while also being economically feasible, both in terms of construction and getting them up there in the first place.

We can make a space data center right now. It would just be terrible and expensive.

reply
The physics problem regarding radiator arrays isn't unsolved, but it's not a problem that scales up gracefully. Small-scale radiators could get by via passive cooling, but large-scale radiators need active cooling, and now you need fluid, pipes, and pumps that all represent additional launch mass and points of failure (and the pumps are generating heat of their own, so now you need more radiators...).
reply
I think it's not necessarily about being the cheapest option, but a more politically acceptable one. I don't think you're going to get people protesting a data center in space considering it won't be next to their house, won't use water, and won't lead to increased electricity rates. I could see companies paying a premium to keep the political heat associated with traditional data centers off their backs.
reply
It's so, so cheap to buy tap water and dump it on the heat exchange.
reply
With well considered engineering it doesn't even need to be tap water. If you have a closed loop thermal conductor that interacts with the components themselves you can then use really trashy contaminated water that just needs to be clean enough not to actively erode the heat transfer mechanism. We have setups like this all the time that use condensed air via cooling towers or salt water immersed heat sinks to discharge energy - it's more expensive than tap water but it isn't technically complex. So if it ever becomes unpalatable (likely due to politics) to use tap water there are some readily available alternatives.

The big win of being in space is just a worse alternative to using an intermediary heat transfer medium.

reply
Isn't the problem also that because of radiation, processors in space either need to have larger feature sizes OR additional shielding / redundancy? Seems like a pretty high price to pay for slightly cheaper energy...
reply
Ars had an article that cited some HPE testing on the international space station that said regular hardware is _probably_ fine up to about 5 years

Definitely not definitive but it's plausible current hardware could survive with minimal modification

reply
hm but "fine" as in probably won't die?

my question was more whether the hardware would need extra redundancy or shielding in order to not have unacceptably high error rates

reply
There's a radiation section in the article, I'll let you draw your own conclusions

https://arstechnica.com/space/2026/07/how-hard-is-it-to-buil...

reply
No. It really is a scam. Everyone with understanding of the physics involved wrote so. Heat dissipation + radiation + launch cost make it a no-go.
reply
Heat dissipation isn't as big a concern as it seems: the weight of the solar panels is significantly larger than that of the heatsinks you'd need per typical modern gpu
reply
The cooling is not an issue for their current designs. Look at their AI1 Satellite specs. They clearly have the cooling figured out. The thing is that it's not a datacenter, it's a single rack. A single rack that weighs multiple tons.

You can figure out the weight of the thing based on the total power output, and "power to weight ratio" from SpaceX's own diagrams. Then look up how much it costs to launch per ton, and even look up what they are projecting it will cost with Starship. Even if they get costs down, it's still astronomical. I just can't figure out who would pay that much money to put a rack into space. There's no way the power savings are worth it. Unless you have some niche where you need your workload in space, I can't see the value at all.

reply
SPCX is valued as an AI company; any and all issues you have with AI company valuations apply to to SPCX.

I too agree that SPCX’s space business is real and valuable, but it’s (almost completely) irrelevant here.

reply
Because SpaceX bought xAI/Twitter.

All of the losses are from the xAI/Twitter side of the house. And Elon Musk needs a flimsy story so that no one sues him. It doesnt have to be a believable story, it just needs to be enough so that no lawyer cares to bring a case in Texas vs SpaceX and Elon for breach of fiduciary duty.

The story did its job. Elon offloaded the money losing Twitter/xAI out of his personal wealth and onto the public through SpaceX. Done and done. SpaceX is now an AI company (or contains one) and needs to perform as such.

-------

It's literally the same story as Tesla/Solar City. Make up bullshit about solar panel synergy with EVs and buy out his cousins failing company. Make it TSLA shareholders problem for figuring out how to make a profit from the failing company, it's no longer Kimball Musks concern since the buyout

reply
The only barely sane rationale that I’ve heard for wanting datacenters in space is that they would give space-x low latency signal processing capacity it would need to turn Starlink into a real time Passive SAR constellation.
reply
> Is there anyone credible who thinks this is a plausible pathway for SpaceX to make huge amounts of profit?

Scott Manly (who I think is credible) has a video where he goes over the logistics of SpaceX's space based data centers. He seems to think its an idea worth pursuing, but its important to note that his expertise is space tech, and not business strategy.

reply
Yeah, I have no doubt that it's possible to put GPU servers in space. The Starlink satellites are servers in space. The question is whether it's even remotely profitable to do this for AI training/inference servers.
reply
The same way that xAI has compute on the ground and there's such demand that the compute exists so it will be sold. I think SpaceX can conceivably sell the space compute easily. This goes with the business strategy of Elon which is just sell services to himself. He sells rocket launches to himself with the excuse of creating a space datacenter. The same way he sells cybertrucks to spacex because the public won't buy them. I'm pretty confident it's definitely cheaper to put the compute in Texas, but might be viable enough to sell himself rocket launches and make his company worth more to leverage for more ventures. I'm not a business person I don't really get how it's all so valuable if the only demand is itself.
reply
Training, no. Latency is too high.

Inference the latency becomes trivial.

Other things, I suspect latency is too high again.

reply
Richard Campbell did a great talk at NDC a month about about this - https://www.youtube.com/watch?v=eo7MEPgWGic
reply
> Data centers submerged in the ocean

Sooner or later it's going to leak.

reply
I think corrosion is the big issue here. perhaps that is where the leak comes from.
reply
Water under high pressure will find any weakness. (For example, the places where the wires go in and out.)

Corrosion indeed is always a problem.

reply
To me it feels like a way for Musk to justify directing AI money towards his first true love, which was space.
reply
The only way it makes sense is if you think the massive job losses due to AI will lead to people burning down datacenters on earth. An older spec'd and unmaintainable datacenter is worth way more than one in ashes.
reply
Datacenters in Antarctica or deep ocean also fulfill that.
reply
Let's just put it this way:

The ISS produces about 120 kilowatts of electricity.

An Nvidia Blackwell B200 GPU uses 1.2 kilowatts of electricity.

So, you would need a similar array of solar panels and radiators just to power 100 of them. You probably would need 2-3 launches for a satellite this big, and realistically, you would just make smaller satellites.

That's $4,000,000 worth of GPUs, A couple millon or more of RAM, SSDs, etc., a radiation-proof satellite housing to support all of that hardware, solar arrays, launch costs ($74M per Falcon launch), all for maintenance to be impossible and the hardware to become obsolete in a couple of years.

It's a delusion unless we invent some way to go to space for free.

reply
Launches are not $74M. That's retail pricing.

SpaceX's launch cost, the internal spend to put one Falcon 9 Starlink payload in orbit, with a return to launch site booster recovery, is about $15M.

If you're going to make such assertions, do the legwork to make sure your numerical claims aren't off by 500%.

reply
I don't really get the obsolete argument.

The thing has two main parts. One, a bunch of solar panels, shielding and radiators. This the heavy / expensive to launch part, but should last for what, decades? Two, a bunch of GPUs. These become obsolete, but so what? They're not that heavy, so every few years you send up another rocket and swap them out.

reply
SpaceX would be launching these on Starship, which has a much lower targeted launch cost.
reply
For reasons already outlined, I have doubts about their math. Targeting $250/kg payload costs is ambitious for a rocket that has not yet successfully reached orbit or proven cost-effective.

Even if we do somehow succeed at affordably dumping tons of GPUs into orbit, what do we do about the Kessler Syndrome?

reply
[dead]
reply