Thermodynamics rules future orbital data centers
spectrum.ieee.org
spectrum.ieee.org
https://andrewmccalip.com/space-datacenters
You can play around with the cost sliders to estimate the economics, but even being quite optimistic, space data centers cost ~2-3x of their terrestrial counterparts.
Heat emission in space is not problematic for humans
If it were up to me, I would be researching and optimizing non-chemical propulsion, also green synthetic fuels for launches. Unfortunately, we will have to leave that for some firm other than SpaceX to develop.
If Nvidia/SpaceX can make the chips run at a higher temperature that would help a lot although I assume it would have already been done if it were possible. Another option is to add a heat pump to raise radiator temperature if the smaller radiator mass can pay for the heat pump mass.
To me the biggest problem of the system is the dependence on launch cost. The more launch cost can be lowered the more viable the system is and the more options there are for solving every problem. But IMO the reason Musk is building this system is to lower launch cost. He needs a project that can fund a high volume of launches and create economies of scale to lower launch cost, the way that Starlink did for Falcon 9.
Starlink doesn't need GPUs, they use hardened chips. The first question is, can whatever works as a 2nm GPU work where Starlink is?
The second question is even funnier: Starlink satellites have a limited lifespan and regularly burn up in the atmosphere when they fall back to Earth - that's due to the need to fly them at low altitudes where, coincidentally, the radiation is lower.
Where then will the flying data centers be - high with higher radiation or low with a higher fall rate? And I doubt those heavy birds are going to burn up fully before hitting somebody's head.
That means that for every 3 units of heat your chips emit, you will use a 4th unit to spin the pump. If your panels generate 150kW, you will only have 113kW available for compute and the rest is cooling. Radiators will more than halve vs 340K operation, so it's net beneficial.
It's all a giant techno-economic optimization problem: the extra mass of solar panels you need vs saved radiators vs mass of the pump, the die temperatures you achieve and corresponding performance, durability and chip price point etc.
If the commoners are attacking infrastructure, there's still always going to be ground stations that can be attacked.
Again, we're getting closer to Hyperion Cantos here.
https://arstechnica.com/space/2025/05/spacex-pushed-sniper-t...
Everybody who knows anything about data centers understands this. You don't even need to get into the thermodynamics of cooling where the only option is radiating away heat. Terrestial data centers have to use water cooling because the heat generation is so significant.
The real problem is failure. When you have a lot of servers, things fail all the time. Hard drives, SSDs, RAM chips, GPUs, motherboard, etc. Servers are designed to be able to quickly replace parts. If your data center is in orbit, these parts will just fail and there is no repair. There's certainly no economic repair.
If I remember my numbers rightly, at a company like Amazon or Google, the ratio is (IIRC) 1 FTE per 10,000 servers, meaning if you have 10,000 servers, 1 FTE will spend their entire time just replacing parts. A good hardware tech will have a pool of known good parts. When a server detects a fault, they'll go and replace probably everything with known good parts and then figure out what's wrong later. In GPU heavy DCs I suspect 1 FTE can cover fewer than that just because the heat management is much more complicated.
I really can't believe we're still talking about this. It borders on journalistic malpractice to even suggest this is a realistic possibility.
We can do it, we just have to decide we want to. If we're going to have a Tulip Mania Redux I'd rather have it with space data centers and reusable rockets and autonomous human-level AI moonshots/boondoggles than... I don't know, literal tulips, or factoring large numbers to mine bitcoin or whatever...
Inference margins are some 70% right now? So it needs quite some work on both sides but this seems to make it seem more achievable than I previously thought.
That's my understanding of what Elon said about how it's going to work.
The difference is that Starlink solves a problem people actually have. Space AI solves very few real problems (immunity from regulation, physical security)
It may do so, initially, compared to a new-build DC.
But, critically, you can upgrade the racks in an already-built DC without demolishing the entire building. You can't do that in orbit, so the full lifecycle ROI is lower.
Orbital DCs are constrained by cooling and by bandwidth. Even a space-to-ground laser (which does not currently exist in a sufficiently-mature form) has a fraction of the bandwidth of a proper terrestrial fiber line. So you're paying at least 10x for essentially a disposable datacenter that can't move as much data in or out, and likely will not be as powerful as a terrestrial DC because of the cooling constraint, just to have it in space because reasons.
I don't see the business case, at all.
This is so transparently another hyperloop-esque pipe dream invented for the sole purpose of inflating SpaceX's valuation.
Mining is also hard and harsh. One of the more hostile environments to operate for machinery. Lot of mechanical wear. And having that in space makes any maintenance extremely harder...
Maybe once we can do all that fully autonomously and by that I mean with near zero non-remote intervention we might start thinking about space...
Nobody wants to build just one, or even just a handful. Musk has talked about 10000 launches a year for deploying and maintaining data centers.
At that scale pollution becomes a serious issue. It hasn't been a problem so far because we haven't launched a lot of rockets. Cumulatively the world has under about 8000 orbital launches and under 40000 launches that were not orbital but reached the stratosphere.
Orbital rocket emissions have an impact far beyond what you might expect just from looking at the mass of what they emit, because they emit much of it the upper atmosphere. Many things that when emitted near the surface are only a local problem become a global problem if you emit them in the upper atmosphere.
If you just want to escape permitting laws, why not float it in international waters? That's a much more accessible and hospitable environment than LEO.
The issue with putting stuff in space isn't the kinetic energy required. In LEO that's about 30 megajoules per kilogram or $5 worth of propellant. The issue is that orbital launch vehicles are not reusable, so you must destroy an expensive rocket to get your payload to orbit. All of these space datacenter efforts are betting that Starship will be fully reusable.
By the time you got permission to lay all the cables you'd need to lay, etc. your competition would already be done with their space data center, or Lunar one in a peak of eternal sunshine or whatever. This is a race to the most compute. Whoever gets the most geniuses in their "country of geniuses in a datacenter" wins.
Wins what? Well, everything they want effectively forever.
Not that it isn't a problem, but I think heat dissipation will have the edge.
so sure, the heat issue seems fatal, but rad-hard designs will certainly have a bottom line impact
Also think it would be crazy to have a worm spread across the servers/starlink, would there be an antivirus system onboard or maybe not applicable, says RTOS can get a virus
I don't know if Musk, or anyone else, has talked about the economics of not being able to (economically) fix things in space. Presumably the idea would be to avoid components that fail more rapidly like mechanical hard drives (use SSDs instead), and to have redundancy so that even if one component fails it doesn't mean the whole satellite/node is now useless.
He must have done the math that at the anticipated failure rate each satellite would pay for itself before failing and needing to be replaced.
It has the killer feature of allowing a human to walk up to a rack and replace a component.
The replacability is nice, but even in terrestrial data centers there are situations where something fails and it never gets replaced, just routed around, until the pod gets ripped out and replaced in its entirety.
Advocating for exacerbating the melting of the polar ice cap, which will endanger dozens of millions of people, just to have more convenient data centers for manufacturing AI slop, is peak HN.
Thanks for inspiring a little fun with the "peak HN" meme:
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
It is assumed that if you want to build lots of data centers up north, you also need to invest in infrastructure. I've seen discussion about smaller modular nuclear power plants, but those things take years. Another thing could be other renewable energy sources.
NIMBY stuff is why we will go to space. In space nobody can tell you “no.”
I think the specific attraction to space is the copious massive amounts of free solar energy, isn't it?
(In reality, they want to build the torment nexus at the Lagrange points because that would just be edgy-as-fuck)
Yes you get more energy harvest from solar above the atmosphere and can orient them to always be pointed towards the sun. But it is still so much more expensive than building out conventional solar in the Sun Belt, which is so much more expensive than just building a massive natural gas plant right next to people's homes.
No, space is desirable because there is no local permitting authority that can push back. People living near data centers have gotten wise to the fact that local people pay most of the externalities of data centers and AI, but the benefits mostly go elsewhere, and the jobs created during construction are temporary. In space, you don't have to lobby/bribe local politicians and astroturf a YIMBY movement.
That's true of just about every industrial, commercial, civic, or residential site. It's the fundamental premise behind every NIMBY protest ever. The benefit of each individual site always runs disproportionately to people further away. It's only in the aggregate, i.e. each individual enjoying the cumulative externalized benefits from far-off, that the equation could ever balance.
To make it worse, underwater tech is notoriously hard to make operationally visible. Sabotage is trivial and undifferentiable from failure and honest error. When we used to work in trading subsea cable cuts in Asia would constantly ruin our best networks. Everyone had point to point microwave expressly because it wasn’t breakable in this way. Exposing compute to this rather than just networking would have doomed the entire enterprise.
He should look at renting space on container ships before considering orbital DCs, IMO. But that doesn't satisfy the critical "Rocket company needs something to do" constraint.
Sounds terrific actually haha. Boy would that be a sight to behold.
0: https://www.datacenterdynamics.com/en/news/nautilus-puts-sto...
There are thousands of container ships in international waters at any given time, so you wouldn't install a few massive power-hungry data centers on a few ships. You'd put a lot of PV-powered data centers on a lot of ships. Pretend you launched them into space, in other words... but don't actually launch them.
And yes, you'd network them via Starlink, just like the satellites would be.
Regardless, every time this guy does something super capital intensive, it looks stupid and then works out for him. So long as it keeps happening, the probability that he is just much smarter than me and predicting the future better dominates the probability that he keeps getting lucky.
Regardless, every time this guy does something super capital intensive, it looks stupid and then works out for him
That's how it looks at first, agreed. But if you look closer, you'll notice something kind of funny. The only times Musk really wins big have been when the competition is either incompetent or lazy (launch services and satellite Internet), or doesn't show up to the game at all (EVs, self-driving and otherwise.)
Data centers and semiconductors aren't like that. Those guys aren't going to stand by passively and let him eat their lunch like General Motors and ULA did.
[1] https://www.theguardian.com/world/2026/jun/09/worlds-first-w...
1: https://www.latitudemedia.com/news/catalyst-building-inferen...
AI inference gives them a great way to turn revenue from cheap electricity without solving the device-to-shore problem yet for that energy.
It merely remains to build one.
With space either you expand even more resources to get stuff back in one piece or it burns... Down on surface you can just ship them to facilities to do this. Maybe even automate lot of parts as these should be standard configurations.
It wasn't an order of magnitude more because of how expensive rocket launches currently are.
(I'm glad that I read the article before arguing this one...)
"However, there are niche applications where the much higher costs of computing in space could be justified. Examples include ... active collision avoidance in the increasingly crowded low Earth orbit."
A self-justifying purpose!
Though there is a simpler explanation: Some guy with lots of stock in Rocket Launches and Datacenters has an extreme financial incentive to push stories where spending on both things simultaneously is the big new thing.
Who thinks they are easy? Elon Musk? The guy that spews obvious (even to him) bullshit like we're going to Mars in the next 5 years?
I don't even want to read the article. It is obvious that Orbital Data Centers have MASSIVE engineering challenges. They may never be cost-effective.