Orbital Data Centers: Why the Hype Outpaces Reality | Hacker News Reader

Orbital Data Centers: Why the Hype Outpaces Reality

spectrum.ieee.org

27 pointsrbanffy63 comments

Data Centers in Space are a practical engineering impossibility, as well as making no economic sense. Engineering and the laws of physics get on the way.

Just because Scott Manley refuses to call that out, so he can do another eight videos about it, don´t stop listening to somebody with the feet on the ground:

"Orbital Data Centers: Spacecraft Constraints and Economic Viability" - https://arxiv.org/abs/2604.27197

"Hot AI in Cold Space: Thermal-Crosstalk-Aware Scheduling for Sustainable Orbital AI Clusters" - https://arxiv.org/abs/2606.26150

"Above the Cloud: Building Data Centers in Space - Richard Campbell - NDC Copenhagen 2026" - https://youtu.be/eo7MEPgWGic

"Space Data Centers Are Dumb" - https://youtu.be/-w6G7VEwNq0

Nobody in this industry with a modicum of applicable expertise believes that orbital data centers make sense financially. Like colonizing Mars and most of Elon's pipe dreams, it's not the stated goal they believe in, it's getting fabulously wealthy from fat government contracts along the way.

Once you understand this about Musk, you realize that everything he is involved with works that way.

How does this get them a government contract?

The US gov't has been leaning into LLMs for a while now (really picked up under this administration) and (the useful and practical) potential of large, cheap (relative to current costs) satellite constellations, particularly for communication and surveillance networks. The LLM pushers in the gov't are letting out lots of contracts for development, either to fund entire data centers or to partially fund them for priority use.

This data center in space stuff is a marriage of the two fields that will likely lead to a lot of contracts and little output, but the not at all secret trick to government contracting is you don't have to successfully deliver to make a lot of money on government contracts, you just have to do well enough to avoid severe penalties and blacklisting.

A demonstration of my point in the last paragraph: https://www.spaceforce.mil/News/Article-Display/Article/4465... - Over $6 billion and nothing to show for it. Raytheon will be fine.

>Data Centers in Space are a practical engineering impossibility,

Not really, thermal radiators are well understood and the size needed aren't really that unreasonable. Radiation hardening the GPUs is probably the single hardest problem along with actual launch costs.

>making no economic sense

Yes this is the real issue, Spacex would need to reduce the cost of launches by ~10x with Starship for it to ever be viable.

I don't think they are impossible, just impractically expensive. That should be no problem for AI companies and their infinite access to capital. Nobody wants datacenters to be built around here on the ground, so I say get to spending.

If it ends up bankrupting them, even better.

the way the crony capitalism economy works, it will bankrupt all of us well before it does the AI companies...

Good thing Elon's companies have a history of moving engineering impossibilities from impossible to slightly late.

Remember when globally competitive electric cars, re-usable boosters, catching a rocket with chopsticks, playing a fps game via a brain implant, and maintaining a satellite constellation at 480km LEO were also impossible?

There's a difference between "infeasible given materials and manufacturing capabilities of the time" and "infeasible as a scalable solution due to the laws of physics"

There is also the very practical question of "what problem is this trying to solve?"

To which the answer for SpaceXaiTesla is always: Hype up gullible investors about the glorious future just around the corner.

*spacexxaiteslacursor

What law of physics do orbital data centers violate? You must have learned some strange physics.

How about the physics of space radiation and radiative cooling.

The guy has to rent out his terrestrial data centers because he can't make use of them.

We probably don't need to pollute space with dead GPU's.

>> What law of physics do orbital data centers violate?

None. That is why they are...in practice... an engineering impossibility :-)

Nobody is disputing you can have a 1kW of compute in orbit...you have that in the space station. But that is a matchbox....

note the scalable in there. Even under best case assumptions, it's only a good fit for edge coprocessing where latency is super important. You're not going to power the world's compute with a self healing Dyson sphere (but if you think that, I've got something to sell you).

Thermodynamics.

Almost no one said those were impossible, just hard. This is completely different. Rather like... a train in a vacuum tube hard. Definitely harder than making a subway with autonomous trains under a modern city. And much harder than being a third rate AI lab, though Elon did hit that target perfectly.

Train-in-a-vacuum-tube isn't even restricted by the laws of physics, and it doesn't have to be a perfect vacuum anyway, just low pressure. Data Centers IN SPAAAAAACE have the teensy little problem of cooling that you can't just brute-force a solution to. Those giant things dangling off the ISS that make up most of its footprint aren't solar panels, they're heatsinks, and they still have issues managing heat.

Well... If the launch cost is low enough you can just pack a radiator large enough. Good thing this might drive the development of a wider Starship and Ultraheavy booster (or a Superheavy-Heavy where three Superheavy rockets boost a bigger, heavier Starship). Eventually we can get to the Comicallyheavy booster.

Perhaps, but I don't see the costs getting anywhere near being worth it. You'd need a space elevator or to manufacture the sinks IN SPAA— er, in space. The latter is theoretically possible but still extremely remote, the former still requires unobtanium.

It all depends on how far you are willing to go to make your server invulnerable to a police raid.

The police will just arrest you and/ or anyone who operates the downlinks or connects to you. The server can stay in space.

"Encouraging" the uplink operator into backdooring the system is also very likely a page in the playbook.

If the user is sophisticated, and has full control over the onboard hardware, this might be difficult. I am sure the satellites that will be built and marketed for this specific use case will be hardened against interference by any third parties.

And, since they are under no real jurisdiction once in orbit, enforcement would be impractical.

Energy radiation scales T^4 so physics is really on your side here. If you can engineer GPUs to run a little hotter you get significant decreases in radiator size required.

That will have to wait 10 years, or even a few more years, for a transition from silicon to a semiconductor with a wider bandgap.

I doubt that, there really wasn't much effort put to allowing GPUs to run hotter. You can always add a heat pump to create the thermal gradient of the GPU can't handle it itself.

Is the starlink constellation profitable on its on without any cross subsidization from any other Musk endeavor?

From the SpaceX S-1: "For launches of our Starlink satellites, the Company does not recognize any inter-segment revenue, rather those launch costs are capitalized in satellites in Property, plant, and equipment, net." In plain terms: SpaceX's rocket division charges outside customers roughly $102 million per Falcon 9 launch — but it charges Starlink $0.

SpaceX appears to be heavily subsidizing Starlink in the launch cost sense.

There is a mix of clever and sketchy accounting going on in Muskworld making it hard to see which elements profitable/sustainable and what isn't.

> Is the starlink constellation profitable on its on without any cross subsidization from any other Musk endeavor?

This is always a funny "gotcha" people bring up. Except the gotcha is always different somehow. Sometimes it's "the US subsidizes it" because NASA gets amazing deals on launches it would otherwise pay 10x or more for. Sometimes it's Starlink that subsidizes the launches. Sometimes it's the launches that subsidizes Starlink.

Maybe the answer is just what is obvious: vertical integration and economies of scale makes starlink/falcon 9 profitable. The combination is the thing.

It would be interesting to see how much of Starlink's business is due to global conflicts, especially the one in Ukraine. If peace should break out, how does that impact Starlink's bottom line? I have strong doubts that rural villagers in the Amazon are a high profit customer base. In my area everyone was clamoring for Starlink to rescue us from DSL. Then fiber came and everyone lost interest in Starlink.

This is not about something being impossible, but something that solves no problem except keeping launchers busy.

The only reason I can imagine for space-based data storage is being out of reach of most police investigators. Recently, a very corrupt banker in Brazil was arrested and his phones were confiscated and, up to now, two have been cracked, with a big effect on the approaching election. If the thing is in orbit, it's a lot harder to confiscate it.

I used to make similar arguments.

But I also now look at actual results, and they are not merely late, they are far less than promised.

Globally competitive EV? Not against the Chinese competitors (Hint: it's why he plans to merge Tesla into SpaceX). Self-driving so good cars would be assets you could run your own self-driving Uber-like service (and customers paid a $10K upcharge for the ticket to do it next year)? NOPE, Tesla can't even do their own self-driving reliably. And whatever happened to the Tesla solar roofs? Can't get one.

Making a car company profitable enough to justify the insane multiples? NOPE, will never happen. But he's selling Tesla now on the Humanoid Robot "vision".

SpaceX did better on self-landing boosters, but they still have serious issues with scaling it up, and competitors are catching up.

But the physics of orbiting data centers make them not absolutely impossible, but very uneconomical. Every problem supposedly solved in space is easier and orders of magnitude cheaper to solve on the ground.

If you haven't yet noticed the pattern, Elon is always selling the next big hype wave. It is always the NEXT thing that will justify the insane multiples. If you want to buy and hype meme stocks on the next-greater-fool theory, good luck; you will do well for a while as there are many fools around.

> you will do well for a while as there are many fools around.

Ideally, you pass your meme stocks to greater fools before they collapse. You then move to the next meme, rinse and repeat. You can do well for a surprisingly long time.

Not being a misanthropist has been expensive to me.

Same here; it's expensive to not be a crook at heart (at least short-run expensive. I think it is far better long-run) .

I understand the point of being able to sleep at night, but lacking a yacht and a business jet of my own, alongside several holiday residences on different tax jurisdictions, is certainly an inconvenience.

This is a very good summary.

Elon hires great engineers. Or at least he did before he blew out his brain with ketamine and went full-on fascist because he had a trans daughter.

He still hasn't built a fully self-driving car. His last electric vehicle (the only one he directly designed) was a massive flop. The successful Tesla vehicles are losing market share because they're old and because of the CEO's odious politics and because he seems to have lost interest in EVs unless they're robotaxis.

Just recently he scammed Wall Street with a horrific overvaluation for a company whose only profits come from Starlink and whose biggest rockets blow up before they reach orbit.

Now he's talking about orbital data centers which make no sense to any engineer who understands thermodynamics (and sadly, many don't) and which are unnecessary because within the next 5 years 90% of inference will be done on local machines in users' homes because the hardware and the algorithms will be good enough to enable it and every Joe and Jane are fed up with surveillance capitalism.

On the plus side, there's enough surplus Tesla equipment that a boutique vehicle factory in Detroit is planning on using some of their cast off engine hubs to produce Kei style truck kits.

have you ever seen a detailed photo of the ISS?

* https://images2.imgbox.com/64/4b/GhtUtq1m_o.jpg

look at how massive the solar panels and more importantly the thermal radiators are outside

compare to the size of the human habitat tubes

"AI" in space would demand as much if not even more for each node

The hotter you are, the less radiators you need for the same heat dissipation. The ISS needs to run at human-comfort temperatures, but a datacenter can run much hotter than that (note the "at 60 °C" in the article). I forget the scaling parameter, but it's like T^2 or T^4 or something like that, so running hotter makes a huge difference.

Of course, that's offset by the enormous heat output of a datacenter. I'm not saying it's viable, just that you actually need to do the math instead of eyeballing the ISS radiators.

I am curious to see if this causes an improvement in radiators though, that would be interesting

or perhaps a whole new technology invented to vent the heat

will be interesting to see what they come up with

hmm, how do they radiate the heat from the capsule with humans going to the moon, now I am really curious, I've never seen panels like that on the lunar module past or future

> how do they radiate the heat from the capsule with humans going to the moon

https://blogs.esa.int/orion/2018/01/18/keeping-cool-while-st... https://ttu-ir.tdl.org/server/api/core/bitstreams/a51a54b9-9...

Looks like the regular exterior of the service module provides enough area for heat exchange.

Rather than reasoning on the feasibility of data centers in space, what are the constraints that prevent them from being built on Earth? What's the minimum cost of building and operating in the US the equivalent of a 1M satellites data center?

The last straw their grasping at is "regulations" are to unfavorable on the ground so space makes sense. People repeat this as though there weren't still a dozen better options than space.

My bet is that all of this "Data centers in space" media campaign is just a cover for testing and launching military stations with an ability to attack ICBM's and ground targets all over the globe.

I am waiting to reserve my place in Elysium. "Get your ticket to that wheel in space while there's time ... The fix is in ... You'll be a witness to that game of chance in the sky"

The advantage is solar power in space. They will never stop getting sunlight.

There's a very narrow band of sun-synchronous orbit with consistent enough sunlight to make orbital data centers viable. Even with a tight constellation of 500,000 to 1,000,000 satellites packed into that band, it still doesn't make economic sense compared to what terrestrial data centers can do. At 500,000 satellites you're looking at maybe 50 GW of solar power. Then factor in the lifespan of LEO at ~5 years you're looking at hundreds of satellite launches daily to maintain that infrastructure.

Now work is being done for increasing solar access (Starcatcher) and others are working on improving LEO refueling and repair capabilities, but I would say we're decade+ away from establishing any true compute infrastructure in orbit.

decade+ is not really that long, is it? Say you're correct in that time frame. If you put money in now to a venture that seems uncertain then in 5 years it's going to seem much less uncertain - and so the investment will be worth that much more.

I think there's an oft repeated view that markets only look to the next quarter. That's not really true. Markets (or people) are willing to go in on long term investments as long as they believe they understand the uncertainty. Which leads to crazy 2nd and 3rd order speculation (I think you think that they think...)

Personally putting money into space based datacenters seems crazy to me currently but in a couple of years it'll be more clear as to just how crazy this is. At which point the opportunity (and risk) will be gone one way or the other.

You're right on that perspective. I should clarify. I mean decades+ when compute and power begins to realistically make sense where we can start planning such a long term strategy. I personally think our cislunar / moonbase efforts are a better effort to put energy and resources into.

In LEO you are about 50% of the time in the shadow of the Earth. Higher orbits or solar orbit (where you can get uninterrupted sunlight) require a lot more fuel and the further out the data center is, the higher the latency.

Solar and batteries get you through the entire day right here on Earth, plus you can shit in a toilet when you need to go on site to fix or upgrade something, not in a little tube. Plus you can probably just drive there instead of needing 15,000 feet/sec delta-v to get to orbit. And you can breathe. Surely the cost of batteries and panels do not exceed the costs and logistical pain-in-the-dick going into orbit will impose.

There is a 150 KW Solar Array on Space X's first Data Center In Space It was launched in June of 2026. Technology improves at a steady pace. Space solar power is new and will improve over time because it is not a mature technology. Also, in space solar power is continuous and receives more energy than on the ground. https://finance.yahoo.com/sectors/technology/article/spacex-...

https://youtu.be/-w6G7VEwNq0?t=340

This has little to do with rationality. It is the classic Move And Fast Break Things combined with the off base idea of building a Kardashev scale civilization. This part of Elon Musk's philosophy as well as many of the tech bros who believe in Abundance. https://x.com/teslaownersSV/status/2068178062030803390

I think everyone misses the real point - they're completely outside the jurisdiction of any government.

Short of launching an ASAT and the risk of space debris, nobody will be able to shut them down.

On other hand anyone using them for anything is certainly under some jurisdiction. Just because no one can get there physically. Does not mean people who use them are out of reach.

> nobody will be able to shut them down.

ASAT can completely destroy it very easily. Then isn't that everyone who can send orbital object can shut them down?

The only countries that can shoot down a satellite are United States, Russia, China, and India.

Is the US going to launch a multimillion dollar multistage aegis missile to blow up a server that is serving illegal gambling?

What are they going to do in these orbiting datacenters that would be getting them shut down on earth?

crypto scams, selling CSAM

I don't think sovereignty works that way.

Reply on news.ycombinator.com