Why put data centers in space?
I am just honestly confused. There seem to be no logical reason for it, except that it would raise the stock value of rocket producers like SpaceX and Blue Origin.
Syndicated from the fediverse. Read and engage on the original instance.
View original on mander.xyzView original on mander.xyz
163 replies
This idea is a total scam. This is just a way of asking investors for an enormous amount of money, with literally no intention of actually following through with it.
They might throw a couple of cheap satellites in orbit, just to say they "did it"...but no one with any actual comprehension of these things is honestly expecting it to yield results. The vast majority of that investment money, is going to be used for unrelated things.
In any sane world, this would be considered fraud. But, here we are.
Remember when Microsoft dumped containers of servers in the ocean to test things.
If they couldn't make that work they obviously can't make space work.
It was always a scam.
You're talking about project Natick. All reports I saw was that it was actually a resounding success. The only reason I've seen for why it was cancelled is because it wasn't able to be kept at the bleeding edge as they couldn't upgrade the servers randomly. It doesn't really compare to launching them into space which gives basically zero benefit but adds an insane amount to the cost.
Not only upgrading hardware, but maintaining any issues along the way. Hard drives fail, Memory fails, etc. That's just stuff that happens and can be easily remedied in a normal data center.
Now take all of those issues, and instead of being underwater... move it 500 miles into the vacuum of space. And take the heat dissipation of all that water, and make it a fraction because the vacuum of space is terrible at radiating heat despite the ambient temperature being so cold.
Cosmic radiation causes random bit flips. You would have to shield everything and make it redundant. More power, more heat, more cooling, more power... Welcome to the loop.
Yeah... No maintenance... Which is a thing that datacenters require.
Enterprise Hardware can have redundancy but things still break.
If it was a success they would have done it. Ergo, not a success.
IDK about success.
I think it went predictably in that the cooling was ok but maintenance was an issue.
Seawater is a real bitch. If you could cool things with it we would just pump it onshore.
"resounding success" but given up as a bad idea.
The fact that it could possibly "work" was not really in question, it was whether it would be worth the costs and drawbacks. It was not.
But either way you are right, space even worse than water with no benefit compared to ocean.
The core reason space datacenters are uneconomical, is the 5 year replacement cycle. Includes trashing the solar and radiators every 5 years even if they could last 30+ because datacenter modules need to auto-deploy their own power and radiators. Ocean deployments would be far better. Wind/solar/batteries doesn't need frequent change. Just drop in a new module every 5 years, while taking out the old one, and rebuilding its insides.
Nah nah, we aren't allowed to do wind and solar anymore. The US president is paying a shitload of money to a lot of companies to stop them from developing renewables in the US and abroad.
The microsoft thing worked. It also highlighted a lot of issues that would also exists in space, but way worse.
But "genius billionaire buffoon" can't read apparently.
It functioned, mostly, but I wouldn't say it worked.
Microsoft? Hahaha. Like Microsoft would know shit about doing anything in space?
While Elon has no issues crashing and burning, no matter how many times it takes, until he succeeds.... That's a true commitment to trial and error!
I wish you could make money off of nowing when a company is spouting bullshit, but after watching Tesla stock for a decade...
I believe Apartheid Pedo cited space being cold as a reason (ie don't have to worry about cooling the servers), except as usual he's a fuckwit with no idea what he's talking about. Space is empty, which means there's nothing to carry heat away, so the only way to lose heat is via radiation. And that takes forever.
Yeah I think space stations have to literally worry about not radiating enough heat away. If there is no medium to transfer the heat into the heat can't really go away. One of the many unintuitive features of space
To the point that they occasionally have to use the solar panels to shade the radiators.
Thermals in space are difficult
We can just terraform space duh
Falcon Heavy with a shitload of ice cubes
There's no reason, Musk is just bullshiting like with Mars colony by 2025.
This. I don’t know how any supports is companies. This man has the attention span of a cocker spaniel on Ketamine.
He just says stupid shit all the time and people are supposed to pretend this it’s a wise business decision.
A list of Elon Musk’s terrible business ideas
Elon Musk’s Eight Most Outrageous Ideas
BS is fair, 90% of it. There needed to be an excuse for xAI merger is origin of the BS.
No, it seems like you understand it perfectly.
It's a shameless attempt to loop in their own assets to a free for all money spree. There is literally nothing about it that makes any sense. It would be more efficient to beam laser light back to earth to power a dc than this, and that's a stupid idea too.
It's a terribly inneficient idea, but it sounds futuristic, which is basically what Elon trades on to acquire government subsidies and investor buy-in.
My theory is that it's so they are beyond the reach of protesters, the law, etc. I think that is going to be a very in-demand property for servers going forward.
With the money that it costs to send an entire data center to space you can easily fortify and militarize your data center on earth to the extent that no protester will ever come close?
The law is anyway already largely under their control right? They don't have to go to space to evade the law.
And to the extent that the law still applies to them and constraints then, doesn't it also still apply to the necessary ground-based necessities for space based data centers? As in, if the state wants to shut them down, they can still do so?
They could be configured such that no off switch exists, no manual administration is required, and it's a fully self-sovereign system that can only be shut down by sending a missile to space, and physically accessing it to extract the data or encryption keys it contains is even harder.
Only if your definition of "protester" excludes anyone using weaponized drones, but that misses the point. The current meta for fortifications right now clearly favors the attackers, and that's especially going to be true for a facility requiring large power inputs. That meta no longer applies if your data castle is in space.
Power grid dependence can be a weak point. Democracy can clash with CIA/oligarchy. A big enough mob can overwhelm security forces not particularly ideologically aligned to defending skynet.
Shooting down satellites isn't scifi anymore. Any medium sized country can do it if they want to.
Why don't they already then? Russia could shoot down satellites providing communications to Ukrainian drones for instance, couldn't they? I think it is because doing so triggers a large shitstorm and real geopolitical consequences, if it becomes the norm then no one will get to benefit from any satellites, so a smaller and less influential country is not getting off easy if they do that, same kind of deal as nukes and access to them except probably a little less so. The similarity being that there is a cutoff point below which no one has access, and above which heavy coercion and other manipulation can be imposed as an effective deterrent.
https://www.theguardian.com/world/2026/jul/07/china-missile-test-where-when-timing-response-controversy
https://united24media.com/latest-news/ukraine-quietly-tested-space-launch-capabilities-in-middle-of-war-twice-17852
Mostly because it's expensive and some satellites are in way higher orbit than you want to fire randomly at, and hitting just a few satellites is more of a very very expensive annoyance unless you're doing it to time it with other strikes (and yes there's also escalation risk, but countries without their own satellites won't care much)
It's a good point that providing the security systems of nuclear power plants is cheaper and more job creating than space datacenters. You can also get more uppity when international adversaries nuke your datacenter on US soil as opposed to space.
But more money to a guy that recycles $100m+ into politicians every election cycle, provides decision maker influence to subsidize space skynet. Continuity of nazi/zionazi ideology may require being beyond any temporary loss of control over law.
They're harder for us to burn to the fucking ground if we can't reach them.
The tech billiionaires got annoyed by regulations and complaints and are looking for a easier way.
We towed it outside the environment.
Into another environment, no?
Meanwhile China is putting them underwater for passive cooling.
They do? MS tried too, but too much issues with marine life, corrosion and accessibility for repairs/replacement. With the third also being a issue with space, but x100.
https://www.theguardian.com/world/2026/jun/09/worlds-first-wind-powered-underwater-datacentre-starts-operating-in-china
Saw this awhile back and kinda rolled my eyes. I see the appeal of the idea, but the execution seems unnecessarily complicated for little gain. The space one is even worse, the sheer amount of fuel to lift that kind of weight is just a stupid waste.
And then they wonder about the spike in climate warming, with water vapor being one of the most effective greenhouse gases.
It was an interesting experiment for certain but it'd be easier overall I think to make a desalinization plant to produce more water specifically for the data center to cool with.
Snake oil. As almost everything else.
You answered half of your own question.
The other thing is a complete lack of jurisdiction. At least theoretically, nobody can go after them for generating grossly illegal content in space.
Also, pure hypermasculine fetishism.
This is the weirdest take on data centers ever.
but ... it checks out?
Because it sounds like "future" to stupids.
I feel like "data centers in spaaaaaaaace!" is an idea that could only come from someone completely removed from the day-to-day of how computers actually work. Like no matter how you slice it, these are physical objects that are going to need servicing by human beings fairly regularly. Putting them in space vastly complicates that process.
They can't build datacenters on earth on time and in budget, it's totally gunna be easier in orbit
Robots exist. They literally have robots that you can basically sync them to your movements. And you can control them with VR. So somebody on Earth can control a robot and fix that shit from Earth, if truly needed. I really doubt they would have any need for that though.
The plan is use modules just like starlink only they're going to call them starmind, and the goal is a constellation of over a million of them. They'll be in relatively low earth orbit and they will have relatively short lives and there will be a cadence to regularly replace them.
I'm sure you've seen how Elon likes to piss away money...
As stated elsewhere it's a scam but their stated reason is because they want to build massive data centres on earth but there's a huge backlog in getting electricity for them. Even generating their own electricity is troublesome because every single natural gas/methane turbine is already spoken for years into the future and solar etc takes a lot of land which they may have trouble getting the rights to build on. Thus, if you could build in space you can theoretically collect a lot more solar and more strongly than what you get on the planet's surface. It just ignores all the other costly parts but I guess if you are being given unlimited money by venture capitalists what else are you going to promise them when they want more more more AI infrastructure
Solar on deserts, or even Nebraska farmland, is far more profitable than corn for ethanol and produces far more energy. Even with 2-5 hours of sunlight/day, battery systems, would be 5x cheaper in Nebraska than space for 24/7/365 power production, with perhaps minor throtling in winter. A datacenter in the middle of a solar farm also eliminates noise complaints from cooling systems.
Can't raid Elons CP database if its floating in space.
We're not even raiding CP factories down on Earth, no need to go in space for that.
You can shoot it down...
They want to see if AI can still function while getting randomised bitflips?
They were already behaving as if they were having random bitflips, how much worse could they get?
Solar power and low latency laser communications.
Power is free* in space.
Communications between satellites is as fast as fast can be, fast enough that it can be faster to go earth-sat-sat-sat-earth than via fibre under the ocean.
The absolute con is that heat is really really hard to shed in space and it's really really expensive to get big heavy things to space.
The entire convo in one sentence.
Honestly I feel like a big issue is the physical security. Now your data center is more vulnerable to any orbit capable adversary.
Lasers are no longer low latency when distances are huge
Power may be accessible with solar panels, but power density is shit in space because solar panels force a flat design with limited depth, plus cooling issues, and you eventually need a construction so large that it's incredibly fragile.
Lasers in space are speed of light in a vacuum which it considerably faster than speed of light in a fibre.
And then promptly loses out to the longer distance between nodes, slower network gear on the satellites themselves, the massive latency loss to requiring more complex higher redundancy signal coding due to the energy loss in the laser across the distance (since it spreads out, vs a fiber maintaining most energy). Even if you used dense LEO satellites like Starlink it remains unreliable and the motion of satellites means you can't have a stable path length.
Tldr it's useless for almost everything. You need to be able to not just use efficient live path selection when sending traffic, you have to be able afford to pay for QoS priority on wildly expensive satellites due to all the specifical equipment and hardware capacity limits. If you're not doing high speed trading arbitrage you can't make that money back.
Just look at what the experts say in the paper below. Below 3000 km it's always slower. Starlink doesn't have the capacity to beat the 3000 km threshold now. USA is something like 4500 km across. The investment alone to beat fiber from coast to coast (the ideal scenario) makes it not worth trying.
https://arxiv.org/pdf/2301.05285
Exactly. It is lower latency.
You're moving the goal posts now to the problem being commercial viability for a subset of activity.
Look at the paper I added. It does lose up to a huge distance because of the uplinks and satellite movement. It takes something ridiculous to get use of space lasers, because otherwise I was completely right, you ADD enough distance to the path for lasers to lose.
And if fiber gets improved further by fibers which slow light less (up from 67% of c) then the threshold distance grows proportionally
Ding ding!
Also, it throws more junk in our orbit, is more difficult to maintain and upgrade, and has a significantly more catastrophic fail result.
But why do things logically or sustainably when you* can make money?
* you = someone who already is hoarding a large portion of our wealth.
But then how do they make money out of it?
By scamming investors and fooling shareholders.
same way musk made a shitton of money with hyperloop: by getting other people to pay for it and then not actually do it.
So I can actually think of a few reasons: Preprocessing vast amounts of super high resolution satellite imagery to save bandwidth, and jurisdictional black site processing data some countries would want to steal or arrest you for having or processing, particularly if you were hiding it from internal audit. You would be paying an extreme premium over ground based for ever FLOP, so, the customer base of practical customers is limited to a handful of very well funded intelligence agencies or scientific projects in very niche use cases.
Like every other real "ai" use case:
This is fucking awful, and I hope there's a hell so every single person involved can get some fraction of the suffering they deserve.
Good news: There is a hell!
If you're gnostic it's right here!
Admire the plug, to call that community hell is really something
How consistent is modern day gnosticism with that of the first/second centuries?
Oh no one of the threads most forms of Gnosticism gave in common is the idea that earth / corporeal existence is hell in the same way that the Buddha says all feeling is suffering. It's like that bit in the good place where she realizes they're actually in hell already.
Interesting! Thank you for sharing!
Another psycho christian who thinks eternal punishment is somehow loving.
Fuck off, death cultist.
I'm the same one you've been following the past few days, lol
Solar energy is more efficient in space because there's no atmosphere and depending on the orbit can be in the sun for much, much longer than on earth over the day. Apparently you can beam the energy down to earth somehow, but you lose something like 50%. So the solution is to put the data center up where you get the power. The downside is that it costs a fuckton to send it up, cooling becomes an issue, temperature gradients become an issue, heat and cold cycles as it goes from sun to the dark side of earth is an issue, and the funnest part is that cosmic rays can fry electronics like GPUs.
You can theoretically beam down energy with microwave… the problem is that it’s hard to concentrate and everything next to your receiver will burn
Seems fitting to have LLMs bring on the age of space death rays tbh
I don't think it makes any sense either:
The only positive is that the energy will be cheap, and that the computers are safe except from nation state actors and untracked space debris.
I lean towards this being a narrative for the markets. Space Exploration Technologies Corp. claimed a 370 billion USD total addressable market (TAM) in space enabled solutions, and a 26.5 trillion USD TAM for AI in their S-1 filing before going public. They state within that their goal over time is to launch 100 gigawatts of compute to space each year.
By making the claim that this would be worthwhile they present themselves as uniquely positioned to profit from this situation. Moreso than other AI compute companies or other rocket launch companies.
Nailed it. Data centers in space is an objectively terrible idea but most people are too stupid to understand why. They see it as free real estate in the ignorance of the logistical problems of constructing structures in space large enough to be of any use, let alone powering and cooling them. The failure of the American education system has made the populace an easy hunting ground for modern Snake Oil salesmen.
Skynet. They want to build skynet. Go watch Terminator. I guess that's going to become a posthumous documentary.
Etc... come on. It's skynet! That's the funny thing you called a data center but really that's kind of the least of what's going on.
Because they're money for space projects and some billionaire wants it.
Right now, it's a terribly dumb idea. Here's why:
For me, this looks exactly like the "we're going to Mars, so invest in our space company, yipeeeee" thing.
That being said, could we see space based computation in the future? Sure. Again, here's why:
Could we develop systems with crazy efficiency? Sure. Maybe. I don't know. But even if we do, I can't see a case where the Hedonic treadmill wouldn't set in, still causing crazy energy consumption.
Hence, in conclusion, boo to current space companies grifting space based data centers. But yay to eventually putting all industry off planet.
With the amount of heat and greenhouse gasses released into the atmosphere with launch and eventual re-entry, will you ever get to net positive on that equation though?
I had been wondering lately if using the fossil fuels in space instead of on earth isn't a viable option. It's like a billion years of sunlight trapped inside the earth's crust that's all gotta go somewhere eventually anyway right? As long as it's not fucking shit up in the atmosphere it could be a lot of use elsewhere maybe?
Haha you're forgetting the second part of the reaction: oxygen. You would need to somehow transport oxygen in addition to fossil fuels in space to burn em. You would need to expend a lot more energy in their transportation than what you would get from them.
Hence, this idea has merit only to be used for propulsion (for now). Once we crack fusion (which I'm sure we will within this century), fossil fuels lose this application in space too.
The radiators would not have to be that large.
They just have to be sufficiently hot(T to the 4th power scaling). For the same unit area you would have of incoming solar flux you would have to move heat to a radiator of that exact same unit area to reach about 160° at a mere 70% emissivity. Heat pumps exist that can do that easily.
large radiators argument is extremely poor. Plan is a mesh of small datacenter modules. The inter nvl72 networking speed is irrelevant for inference/skynet. However, there is no hot semiconductor technology theory that competes with silicon for feature size/speed. There are much better designs than circulating fluid radiators, but all of them, add complexity and pump/nozzle failures. Supercritical compression (heat pump) systems is vastly more prone to failure, including in pipe/connection network. It also means thick, less deployable, radiators.
Transferring huge amount of data between space and earth's surface is costly and a big bottleneck. If you have a lot of data coming from space (satellite watches and listen to a lot of stuff) that needs processing, you can hit that bottleneck. Moving processing power into space removes it. Data transfer, using any kind of wave (even light) can be make way more efficient, so you can process a massive amount of data way faster, and only send back down there the result.
This has two obvious type of applications : useful for humanity, and dreadful for humanity. Think, doing weather forecast vs. using worldwide pinpoint tracking of any electronic device, including ones that don't necessarily have wireless capabilities. Guess which one will be prevalent.
Well, that, and doing a pump'n'dump with investors money, since for now the technology to do that reliably isn't really a thing, and spacex whitepaper just goes "yeah, we know this is a problem, we don't care".
The problem starts when you try to do the math for power and cooling. Plus they're relying on the deployment being almost free.
And then there's the maintenance...
This guy gets it. Musk is building skynet.
There is zero economic reason. Most optimistic cost estimates are 5x more expensive than earth systems powered by solar and batteries. With near term optimism 10x more expensive.
The only reason is skynet that is immune for anyone in the US pulling the plug from skynet. Russia/China can already hit sattelites or terrestrial datacenters with missiles, and space ones are more vulnerable because Americans are more likely to cheer destruction of skynet assets used to oppress them.
The other factor is international law permits squatters rights to the most desirable orbits. 300 years from now, we may need power/data beemed to/from space. Privatizing it for oligarchy is a great long term investment if it is subsidized by US government.
I remember watching a real engineering video, the white paper for it doesn't go into properly explaining the cooling mechanism. There were plenty of other problems too, but this alone was enough for me to know it's running on pure hopium.
Oh and you cannot just use the corporate maths in this and not count initial costs, and costs of assembling the whole thing. Estimated area of solar panels for a datacenter of 5GW was 12.5km² assuming the best of the solar technologies used. The initial cost of putting it up there would be so high, it might even dwarf the lifetime operation cost.
In my opinion, it's just another grift like hyperloop ko keep the AI bubble from popping.
Edit:5 not 4gw
Edit 2: I had watched the video long ago, and I glossed over a lot of things,
https://youtu.be/_qpdUNMt2yg
Here's a link to the video to go watch yourself
That video is very poor, for focusing on a red herring of trying to build a monolith 1gw datacenter in space. There is no plan for such. Instead a mesh of smaller (fits in 1 rocket) datacenters that each have their own solar/radiators auto deployed with the datacluster. Inference does not need fast networking between clusters as models fit in a single cluster easily.
It's still completely uneconomic because even GPUs that survive 5 years, are worthless after 5 years. Plan is to trash the datacenters every 5 years including the solar and radiators.
Many of the issues people have with datacenters appear to be "solved" by moving them to space. A few of the solutions are based on common misconceptions and actually wouldn't work (not to mention all the additional 'cons' to building them in space), but to someone who doesn't know better it might seem appealing
They're noisy - no sound in space, plus they're far away from people
They take up a lot of real estate - lots of space in space (preferred orbits actually tend to be pretty full of stuff)
They require a lot of power - solar power is very efficient in space because no atmosphere
They use a lot of water for cooling - space is cold (it is not)
Exactly. Thanks. My question then is, why would rich people waste their people waste their money on this?
Is there some kind of longer term interest that Im not seeing. Are they actually building towards something that does make sense on the long run or something like that?
ever heard of the Nazi bases on the moon?
yeah. that's why.
there isn't any good reason. you should start giving less attention to those who shout loudest, IMO.
To get space data.
Not space centers?
I don't think we've found the space center. There might not be one. We'd need way more space data to figure it out.
You just wouldn't. The ISS is the largest thing we've ever built in space previously. It has an internal area about equal to a 747 passenger jet. That's about 4800 sq ft, but the entire structure is about the size of a football field, with the solar panels needed to power it and whatnot - that's 57,800 sq ft total space taken up to get that 4800 sq ft. The average data center currently is around 100,000 sq ft. The ISS took about 1000 hours of space walking and 136 space flights to build. So we need 20 times the ISS to build one average data center, so 20,000 hours (that's a team of 10 people working 12 hour shifts for 166 days) of space walking to build, 2700 space flights, and it would^1^ take up over 1 million square feet of space in the sky.
OH, and that is completely ignoring the fact that ISS orbits earth in 90 minutes. Maintaining data transfer capabilities from that to Earth would be a nightmare and building a structure that big that close to Earth also means it's going to be BRIGHT in the night sky (like, ok, not as bright as a full moon or anything, but 20 times brighter than ISS which is the 3rd brightest thing in our night sky already... This data center would be well ahead and into the number 2 spot, you literally wouldn't be able to miss it).
OH OH, that is also ignoring how incredibly difficult it would be to cool. Cooling components on the ISS is one of the biggest challenges because there is nowhere for the heat to go. So add on hauling several million gallons of water to space every month I guess? I dunno what their plan would even look like for cooling that is such a monumental problem.
AND THAT IS JUST ONE AVERAGE DATA CENTER. The cost would be astronomical. The environmental impact would be devastating. The repercussions would be severe. The benefit would be almost non-existent and very short lived if there even was any.
^1^ - Ok, this is all napkin math based on what it took to build the ISS. Could it be done more efficiently the second time around? Maybe, but we aren't just talking about ISS 2: Electric Boogaloo either, we are talking about a giant 20x structure. Chances are it would be even worse than these estimates - and where the hell is SpaceX going to get the actual physical man-power to do the construction of this thing in orbit??
That also assumes this is done in the same orbit as the ISS right? The ISS isn't in a stable orbit, right?. It occasionally needs to adjust its orbit because of atmospheric drag. You're talking about 20x the area which could be a considerable difference in mass. That's more fuel needed per burn which also needs to be transported up to the data center. Otherwise you are looking at putting this thing in an even higher orbit which comes with its own host of problems.
Oh yea, 100%. I meant to talk about that in my first "OH" section, lol. The ISS is low. You either have to build it further out or have a nearly constant supply of fuel to keep it in the sky. Building it further out means more fuel and time to get to it and less frequent viable mission launch windows, so more difficult to build, but that is still probably the smarter choice than building it at the ISS altitude and needing rocket fuel sent up almost as much as water for cooling...
Like, actually building a data center in space would bankrupt Space X. But the hype from a bunch of dumbasses saying they are going to makes them a shit load of money, because our system is dumb.
This idea makes sense only for starlink and similar networks.
The main issue with such satelite internet is ping. It takes few milliseconds for light (transmitting data) to travel to satelite and again back to servers on earth. This communication happens quite a few times for example in case of loading a website. There are probably some usecases where companies/users would be willing to pay for such speed increase.
Polluting sky with servers could reduce latency by half. For people using fibre on earth this does not make sense. We are better off with servers on the ground.
Perhaps an appropriate place for the billionaires to go. Go, fly with the AI, what could go wrong?
Coldnvm I'm stupidIn space, no one can hear you brrrrrrrrrrrrrr
There's a whole lot of answers in here saying cooling is impossible or super hard, and that people whose literal job is building satellites haven't thought of this. Or that the people whose literal job is building satellites haven't though about all sorts of other basic stuff like "you can't do maintenance on satellites" or "there's radiation in space" or whatnot.
Yes, they've thought of all those things. Those aren't useful answers
The main advantages continuous high-power solar energy, different regulations being in play (it's "outside the regulatory environment), and in the case of SpaceX it provides synergy with their own rocket launch business. They're creating their own demand, allowing everything to be cheaper in bulk.
Yes, it's to "raise the stock value." Isn't that why companies do anything? Doing a thing that produces value is going to raise the value of the stock of the company doing it.
That's where you're wrong, this is like the hyperloop or solar roadways, where all of the experts agree it's a dumb idea but you have a rich person trying to sell that to the public with the promise that they'll solve the issues by the time they get to production.
You can't cool things in space effectively, the only way things lose heat in space is through radiation. According to Wikipedia the radiators expel from 100 to 350 W per square meter. My PC generates more heat than that, imagine a data center. A quick Google tells me it's 5-15kW per rack or 40-60kW per AI rack. Let's imagine an AI data center, a single rack produces 50kW, so you need 140 square meters per rack. To put this in perspective the space station could host one rack if we removed everything else that produces heat from inside. So you need a megastructure just to host one rack, that will become obsolete in a couple years.
It's not feasible, it's just as stupid an idea as a vacuum tube for trains, and brought to you by the same person too.
Here's the usual video I use for this issue, a walk-through of the equations on how things cool in space where Scott Manley runs the numbers and shows that it's perfectly reasonable with a modest radiator panel.
What temperature are you assuming for the radiators? At 80° Celsius (353 Kelvin) they can do 880 watts per square meter. And bear in mind that radiator panels are double-sided, so that makes 1760 watts per square meter of dedicated panel. 80 degrees is what these chips are meant to run at, if you wanted to you could use a heat pump to make the radiator panels even hotter and they'd get much more efficient - according to the Stefan–Boltzmann law thermal radiation goes up proportional to the fourth power of the temperature. Shouldn't be necessary here though.
Unless I missed something, all of his math assumes 20kW of computers, that's not even one rack for heavy stuff, of course you can run a data center in space if you send one satellite up for each computer, that's just unpractical.
Let me put this into perspective, what you're saying is like someone has an idea for a revolutionary parking plaza, and they're showing you it's feasible to spend millions of dollars to park one motorcycle in it.
The actual Starmind satellites will have 175kW of compute on them, Scott made that video before the specs were released so he assumed a minimum package. But all of this scales linearly, if you double the power usage you just double the size of the solar panels and double the size of the radiators. If a 20kW satellite works then you can bolt ten of them together and get a 200kW satellite that works too.
Cool, so we moved from a single motorcycle to a couple of cars, still a long way from a parking lot. You're not understanding the issue, the scale is not feasible, you'll spend millions of dollars to launch a single satellite that's not even 1/10th of a tiny data center.
They're not putting up just one satellite. It's going to be a constellation. The constellation as a whole will be equivalent to a data center.
Also, it won't cost them millions of dollars to launch a single satellite. That's why this project is paired up with Starship development, the two go hand in hand. Starship makes the launches cheap, Starmind gives Starship the scale of demand a ship like that needs for its economics to work.
Even by their own estimates it's 2-10 million dollars per launch, imagine thinking that spending 6 million dollars to get a single rack in space where it can't be maintained and will become garbage in a couple of years is a good investment. And that's just the launch price, you still need to build the satellite, with its hundreds of square meters of solar panels and radiators. There's no point arguing this any more than there was a point arguing hyperloop, until they present a feasible product it's just big talk.
Yea. Heat pump it upto 500k? 6.4 kW/m² (double-sided) for 120 kW total output? 20 sq meters. That's nothing.
I suppose one could, They're going with 160 m² radiators (double sided of course) so they won't need heat pumps, though, I guess they found that to be a better tradeoff of mass and complexity.
160 m² makes sense at 80C... That's right on the money.
To protect them from missiles and drones
It’s cold up there
And yet the worst place to cool something since it is a vacuum.
You really don't know shit about radiators do you?
You really don’t know shit about heat dissipation do you?
https://youtube.com/shorts/HHY6ju0kJ5g?is=0suw8lP7WyDyZ9Iw Here it explains why it won't work.
Edit: And that is only cooling, we are not even talking about the production of electricity and the fact that processor is space a limited in how small they can get due to solar radiation.
Why put trains underground? The question isn't of practicality, but if the idea would sound cool to a 10 year old.
Trains underground? What you mean like subways? There are actual reasons why you would want to do that, like making public transport infrastructure in dense urban locations where you can't just demolish buildings to make space for your train tracks.
Im specifically referencing hyperloop. Of course there are some contexts (usually dense urban ones) where trains in tunnels make sense, just like there are some contexts where putting data infrastructure in space makes sense.
Wait, does the hyperloop make sense to you?
No, in the same sense that "putting a data center in space" doesn't make any sense either. If they had decided to put high speed rail above ground in 2013, it could have been done and dusted by now. It could extend the whole coast of California by now. Digging that shit underground? They would be lucky to have 100 miles done. Its completely ignorant of the context, costs and impact associated with its design decisions, something that has become a pattern in the companies this particular manchild collects like playtoys.
Ever wonder why these guys never touch paypal? Because its their perfect cash cow. Its their credit card company double skim off the top scam. If they ran paypal like they ran the any of their other companies or the government, they would destroy it. They know that they are incouragable fuckups incapable of bettering society. Its why they go about scheming to have control of it instead. I'm convinced hyperloop was never anything more than a sales pitch to derail any real progress on a genuine public asset that would cause fewer electric cars to be sold.
Elon Musk's biographer has apparently confirmed exactly that, that all the hyperloop hype was intended to derail investment into high speed rail.
Isaacson is a ball gargler of the first order.
I'm assuming they were thinking of freight trains, possibly specifically in the US, where they have to cross thousands of miles. That would in fact be a massive waste of tunnel digging.
"trains underground" are Musk's "HyperLoop"
"data centers in space" are also from him.
Yes, Elon Musk projects are most often bad.
ummm....As a ex-Londoner, wat?
Look up Hyperloop and The Boring Company.
Well, they'd be solar powered. Wouldn't contribute to climate change, etc. no unlicensed methane burning turbines in Memphis or other places around the country causing asthma for little children and families.
Ideally if we could export all industry to the moon (so it doesn't eventually crash into the Earth) that'd be really great, too. Solar would be excellent there. Lots of good stuff to mine...
Ultimately though by data centers they probably mean things that are going to surveil what's on planet Earth and yeah... And you know then the people who don't like surveillance will have a really hard time to get to those data centers.
I think your problem is you just kind of like stuck in your own point of view and maybe you need to see things from other people's point of views? Even if those people want to explain you surveil you know everything you do we know everything you look at online every single last image every single last week every single ass porn video.... ❤️
The rocket fuel to get them there, plus all the energy to build and produce the rocket fuel definitely is a contribution. They could be solar powered down here. The reason they’re not is money.
Plus the ozone layer destruction as they re-enter the atmosphere when decommissioned
No no, we don't talk about the vaporized metals in the high atmosphere...
Funny how you peeps never talk about cooling those space slop generators eh?
Radiators work in space. The area required for solar arrays easily accommodates cooling. And there's the issue that you actually need to heat electronics in space for them to function properly.
So really the issue is heating them and not cooling them. But yeah overall you would need heat pumps.
It's funny how a guy with no engineering experience whatsoever thinks they know more than people putting fucking rockets into space.
You can even do fun things with TEG. Utilizing the difference in temperature between two surfaces to generate energy.
And the hotter your radiator the better it radiates. So yes heat pumps would move heat, which is funny we call it heat cuz it would actually be relatively cold compared to the radiators (potentially), and your radiator operates at T^4 scaling. You get that thing(radiator) as hot as possible that would be the goal.
Really it's more of a Goldilocks problem. Or basically just a heat pump problem you have to be able to pipe your heat perhaps to the processor perhaps away from the processor depending the Tj(thermal junction.)
That is just a load of bs. Heat pumps do not operate efficiently at high temperatures, and they only add more heat to the circuit. You cannot generate energy from the temperature difference, because then you would slow heat transfer significantly while generating an insignificant amount of power.
It has already been calculated, and radiators are a really big issue for "space datacenters". Even with our current cooling technology advancements. Server hardware needs to be kept at a cool temperature (ideally not higher than 50°C), and most importantly, it needs to be constant in spite of very variable hardware load. That is really difficult to manage in space, and an issue that needs to be considered even for our non datacenter satellites.
But even if cooling was trivial, space datacenters make absolutely no sense.
For starters, it wouldn't be possible to put any server in the market up there. Chips are very vulnerable to the radiation in space, which is why space CPUs are stupidly expensive and way slower than earth CPUs. They need to use less efficient litographies, with very special shielding that isn't commonly available. The sheer compute and memory density of our current servers would be impossible to properly shield, even if we ignore the cost. ECC RAM and ZFS pools for drives would very quickly stop protecting your server from errors, since the bit flips would be too common, and your computations would often be incorrect.
If you ignore both cooling and space radiation, you still have the issue of accesibility. Datacenters have people always coming and going, because hardware needs maintenance. Sometimes you want to upgrade the server, change the connections, add or remove hardware, and of course, you need to replace the parts that break over time. Hardware becomes old really quickly, especially for the compute intensive workloads that are planned for those "space datacenters".
Datacenters are kept fast and competitive through constant upgrading of the parts. Compute intensive servers also churn through parts very quickly, because even at 50-60°C temps, keeping hardware at 100% utilization isn't the best. This is further exhacerbated by temperature fluctuations, which can prematurely kill the silicon.
Like someone else mentioned in another comment, Microsoft already tried dumping containers with servers into the ocean, and putting wind turbines on top. That's free cooling, very stable temperatures, and free energy. Even if it was a success in every measure, Microsoft dropped it because it wasn't accessible enough, which made it unusable for any real usecase.
The truth is, datacenters are a market where cost is very important, and solutions are already very competitive. Putting a datacenter in orbit would be stupidly costly, even if you could fix all the issues mentioned. Free electricity does not offset those costs, and that's without taking into account the cost of the rockets for putting the datacenters in orbit.
TL;DR: There are so many issues and extra costs, even when ignoring cooling, that make datacenters in space a stupid thing to even consider.
And before you say I'm not an engineer and I don't know what I'm talking about, server infrastructure is my field. I've worked with datacenters, maintained server stacks and I myself have worked on those cost calculations. I know what I'm talking about.
Thanks for saving me the writeup. Probably would have taken me some research to find peer reviewed sources again but this is exactly what I have found in the past.
Plus the engineers they rave about are the same ones incapable of building a new giant rocket with a launch capacity to orbit over the one banana threshold so that made me chuckle.
https://youtu.be/FlQYU3m1e80
Grab some popcorn and learn something.
Opens link, sees Scott Manley and realize we ain't in peer reviewed territory by a long shot.
This is a troll, they used chatGPT (which they called "Chad GPT") to reply to me before. They don't know shit, but it's funny seeing them think they do. Read their user description, it's so funny.
You can't even understand the Stefan-Boltzmann equation. Gotcha. Have a blissful day.
I think your problem is that you are thinking of typical datacenters with commodity hardware orbiting the planet. Which is obviously Ralph-levels of ridiculous.... Mmm paste!
Anywho, for entertainment I'll respond.
Specialized heat pumps can hit several hundred degrees. Recall T^4(Stefan boltzmann law)? 120C for a perfect blackbody can balance the incoming solar flux at 1 AU. Realistically, a coefficient of emissivity(epsilon) of 0.7 would need a radiator at ~160°. This is trivial. There is no heat generated in the circuit unless there's a direct energy loss creating it ("friction"). It's conserved and it doesn't matter. So what?
Semiconductors don't have to run under 50° or stay "cool" nor is there need for any sort of constant temperature. I have never heard anything so misguided. That's not even true for terrestrial data centers. There is an extremely wide range of operational temperatures for hardened semiconductors built for purpose for orbital deployment. And since that's what we're talking about maybe that's what you should think about? Moving on...
Take xAI/SpaceX, they plan to modularize them into 100-150 kW satellites that'll mesh network. So if you get 7,000 to 10,000 of them you then have a 'gigawatt datacenter.' Not one giant monolithic multiple square kilometer datacenter complex (which seems to be what you're thinking.) oof, moving on...
Accessibility? Deorbit and launch upgraded replacements, lol. Which addresses quite a few points... Just like starlink. What the fuck is the story about Microsoft? So what?! How is it even relevant? I think you've been on Earth too long. Moving on!
Technically, you would have custom Asics built for purpose... You may have heard of terafab? At least that's the plan for Elon muskrat! You can trade-off shielding mass, radiation hardening, redundancy, error correction, orbital altitude, component selection, replacement cadence, and so forth. You know, specifically design it for the environment. Does that make sense to you? Are you groking it yet, sunny? ;)
Honestly, I'm totally skeptical about the economics, but tech bros have a shit ton of money to throw at the problem. Do I think we should do it? No, but when has that ever stopped anybody with more money than sense?
You don't know what you're talking about. Heat pump circuits don't generate extra heat? Semiconductors don't need to stay cool? They don't need constant temperature? You don't need accesibility? Dismissing Microsoft's story because it happened on earth? Quoting Musk's numbers without questioning if they're even possible? You think ASICs will fix everything?
Yes, heat pumps generate heat.
Yes, semiconductors need to stay cool and have constant temperature.
Yes, you don't need to use standard hardware, but then, why spend 100x more on hardware to have 40x less performance? It makes no sense. You can only use appropriate hardware on earth.
Yes, you need accessibility. Are you gonna deorbit and burn a whole "100-150 kW" satellite just because you needed to change a cable on a patch panel? Microsoft's story demonstrates how even with free power and cooling, accessibility breaks the deal because it's that important. And Microsoft "only" had to refloat a container and open a hatch. They didn't have to burn the whole thing.
Additionally, burning metals in the atmosphere is incredibly harmful for the ozone layer. So there's that.
Even if you built that "gigawatt datacenter" in space, that's a meaningless term. It only makes some sort of sense on earth, where we know the performance/watt of common hardware. The term is a sensationalized thing, and the correct measure is performance, like flops. A gigawatt datacenter in space, with hardened hardware, will be way slower than a 50 megawatt datacenter on earth.
You say you could have custom ASICs built for the purpose? While some massive companies have made their own hardware before, look at nvidia. The only reason they're at the position they're in is because they have designs for AI accelerators. Only nvidia can design those fast chips, and even if they're really expensive, most AI datacenters (and even compute intensive non AI datacenters) are full of those. Nobody can build anything, ASIC or not, that beats nvidia.
ASICs are also a really bad choice for a datacenter, because they're expensive, have to be custom built for a task, and make your super expensive datacenter useless for everything else. Even changing the training method of the AI will make your ASICs useless.
And no, you can't just "build a chip for the environment" and expect to not have to cool it to the same strict levels, or shield it, or have error correction. Semiconductors are not magic, they work in a certain way. And of the litographic processes announced for the so called "terafab", none of them can be radiation hardened. They're too small.
I don't know why I even bothered to write this, and I don't think I'll bother to write the next reply when you inevitably quote more of elon's delusions at me. You clearly have never been close to a datacenter, and you don't understand what you're talking about. You believed musk when he said it totally made sense, and your whole argument makes no sense at any level.
I didn't quote Elon at all. Why you so upset?
You're arguing against a version of the idea that nobody actually proposed. And you're still doing it.
ELI5:
A computer uses electricity. Almost all of that electricity eventually becomes heat. In space, you don't cool the computer by blowing air over it. You conduct the heat to a radiator, and the radiator emits infrared radiation into space. That's literally what spacecraft have been doing for decades.
A heat pump doesn't magically destroy heat, either. It moves heat from one place to another while consuming additional power. If you have 100 kW of computers and a heat pump consumes 10 kW, you ultimately have roughly 110 kW that has to leave the spacecraft. Nobody is disputing that.
The important question is whether 100-150 kW can be radiated into space. It can. At roughly 400 K, an idealized high-emissivity radiator emits about 1.3 kW/m². So 150 kW requires on the order of 115 m² of radiating surface(they use double-sided radiators so it's actually half that), before accounting for engineering margins, Earth IR, view factors, etc. At higher radiator temperatures the required area gets dramatically smaller. This isn't speculative physics.
"Semiconductors need to stay cool" is also not equivalent to "semiconductors must be operated at terrestrial datacenter temperatures." Semiconductor reliability generally improves at lower temperature, but the allowable junction temperature is much higher than room temperature. Different electronics can also be designed for different temperature ranges. The spacecraft thermal system's job is to keep the junctions within their specified limits, not to maintain the entire satellite at 21°C because that's what a conventional server room does.
Radiation is a much more legitimate engineering problem. But "the chips must therefore be old, slow rad-hard CPUs" doesn't follow. There are several ways to address radiation: shielding, redundancy, ECC, watchdogs, scrubbing, fault-tolerant architectures, selective hardening, shorter mission life, and using commercial silicon where the risk is acceptable. Starlink already demonstrates that SpaceX is willing to use this general philosophy rather than designing every component like a 1980s NASA deep-space probe.
And ASICs aren't supposed to "fix everything." They're useful because AI/datacenter workloads are already extremely amenable to specialized accelerators. An ASIC doesn't have to be a single-purpose chip that becomes worthless if you change one training algorithm. Modern accelerators contain programmable processors, memory systems, matrix engines, interconnects, etc. Even Nvidia's own products are effectively highly specialized compute architectures rather than magical general-purpose computers.
"Nvidia is the only company capable of making fast AI chips" is an assertion, not an engineering law. Google has TPUs. Amazon has Trainium/Inferentia. Microsoft has Maia. Meta has MTIA. Jim Keller is behind tenstorrent. Various other companies have developed AI accelerators. The existence of Nvidia does not establish that nobody else can design an accelerator.
Likewise, "100× more expensive and 40× slower" is a number you invented unless you have an actual orbital hardware architecture and benchmark to support it. If the claim is that orbital compute is uneconomical, then compare actual quantities: $/FLOP, FLOP/W, useful compute per kilogram, launch cost/kg, radiator mass per kW, solar-array mass per kW, radiation-induced error rate, expected lifetime, and maintenance/replacement cost. That's an engineering argument. "It can't possibly work because terrestrial datacenters don't do it" isn't. And the Microsoft underwater-datacenter experiment doesn't demonstrate that orbital datacenters are impossible. It demonstrated that maintenance accessibility has value. That's obvious. An orbital data center would have to be designed around that constraint, just as satellites, submarines, aircraft, offshore platforms, and nuclear reactors are.
Finally, the "gigawatt" point is fair in one narrow sense: watts alone don't tell you useful computational performance. You need performance/W, performance/kg, cost/FLOP, etc. But that doesn't make gigawatts meaningless. Power is one of the fundamental constraints on a compute system. If somebody puts 1 GW of useful compute power in orbit, the relevant question is exactly what performance they get from it. So yes, there are serious engineering questions about orbital AI datacenters. Radiation, thermal design, mass, launch cadence, degradation, optical communications, power generation, and maintenance are all real problems.
But "there are engineering problems" and "therefore the concept violates physics" are two very different conclusions.
The former is engineering.
The latter is just internet physics.
You just got fucking schooled by Chad GPT. Have a blissful day.
I said I wouldn't answer, but I just can't help it. I'm literally laughing at this reply.
Yeah, I guess that explains why everything is so misinformed and incorrect.
I'm not even gonna bother answering any of the points, as just 5 minutes of research is enough to disprove them.
Though, I am gonna share my favourite part of your rant:
Yeah, I guess the Application Specific Integrated Circuit doesn't have to be application specific after all.
I'll also share this other gem with you. You wrote it yourself! I just made a few changes to it to make it more accurate.
"It's funny how a guy with no engineering experience whatsoever thinks they know more than people designing server systems and computing solutions, who have experience working with datacenters."
Direct solar, easy heat dissipation, SpaceX stock price
"easy heat dissipation" lol there is nothing easy about getting rid of heat in space.
You can just wait but it's slow
Overheating is a massive issue outside the atmosphere. There's no passive airflow to transfer heat through convection and there's no air at all to utilize conduction. The only method left is radiation - the mechanism that let's you feel the heat of a fire from the side. That's the slowest option of them all.
Space is not cold in any sense of thermal management because space is essentially empty. Representations of people/fluids freezing when evacuated into space isn't a good representation of actual thermal management. While there is some freezing effect, it's because the rapid vaporizaiton of water cools the remaining water and can go below the freezing point. But, realistically, bodies don't really freeze to an appreciable degree. Meanwhile, for everything else not containing water, there's nothing to cool it down. In fact, of the very few particles that are flying around, they're actually quite hot. Atoms flying for millions of years are still over 20,000 degrees.
So no. Thermal management makes orbital data centers the actual worst option.
The heat radiation is actually significant, but still stupid for data centers (high power density).
Those particles flying around are very easily many many orders of magnitude hotter than 20'000 K/°C/°F. To convert average kinetic energy to temperature via the kinetic theory of gases: T = 2E / (3 * k_B) k_B = 8.6E-5 eV/K So roughly T = 7736E Example with 1 TeV: 7736 * 1 TeV (1E12) = 7.7E15 Kelvin
Even solar wind is ~1'000 eV and thus well above 1 million Kelvin. Wiki has a nice graph of amount of particles vs. their energy here: https://en.wikipedia.org/wiki/Cosmic_ray
1 TeV is almost completely to the left of that plot and there is about 1 such particle every second per m². At the right side, around once per km² per year, with 1E19 eV the "temperature" is about 1E23 Kelvin. One(!!!) such particle has dozenz of Joules in energy.