I've been hearing a lot of talk about using satellites for data processing lately. The benefit is better energy generation as I have understood it. But isn't there a problem in that satelites are bad at expelling excess heat, since they can only do so passively?
heat dissipation on satellites is a cataclysmic showstopper, but even if it was solved, anything on a satellite has to survive launch, is nonrepairable and all of that expensive silicon burns down in a couple of years
"datacenters in space" story is a result of highly advanced reasoning: 1. musk has space launch company 2. musk has ai company cash furnace 3. musk wants more money and 4. for what i assume are completely rational reasons and not speculation fueled by cultists, ai stocks pumped and only started crashing recently
Nah, man. I have an issue with your wording. You have not been "hearing a lot of talk". You are quoting one specific lie of a prolific grifter.
Nobody in their right mind agrees with this. If you had put your question as "Elon Musk, owner of a satellite launching company, as well as several money burning pits, wants others to pay him to put some of those money burners into orbit. Is there any sense to this?", the answer would be trivial.
Absolutely. Data centers on earth are subject to the laws of physics. Data centers in space, per SpaceX's white paper on the subject, operate beyond paltry thermodynamics, material sciences, and several other fields of science. When you're that much of an IPO boosting gigachadalphabro it's all profit.
part failures (with so many parts, there’s always something broken
eol parts (lifecycling in 3-5 yrs avg)
They are all challenges that can be overcome. But at what cost. Consensus is that economically it seems hilariously unviable. Datacenters tend to have to make their money back within about 3 years.
The 3 year make money back issue is also why I believe the AI DC buildout is such a high risk gamble/bubble. In a year or two, all that hardware they made a year ago will need to be replaced. You just redo the whole investment again…
No, they are not. Putting a data center in space is a completely ridiculous idea. You can probably build and run hundreds of data centers on earth for the price of one in orbit.
I'd really like to know who put this utter nonsense idea in circulation. Wouldn't be surprized if it was Elons idea...
It would be astronomically expensive, wasteful, and expensive (yes twice).
Got a server rack damaged? Are the IT going up to fix it or astronauts? What kind of training would there need to be? What about upgrades? What about waste or recycling of old equipment? Do you just eject out into space wasting all the precious resources??
It's a completely stupid thought that they hoped sounded cool, but it's not really beneficial in any way
What they don't show in a lot of SciFi is the mundane day-to-day stuff (AFAIK, the only shows that came close were Red Dwarf and Firefly.)
Problem is, things break. Connectors come loose. Temperature makes things brittle. Small debris cross pads and create shorts. Most automated manufacturing lines can catch assembly problems, but none have 0% failure rate. This is what post-deployment maintenance is for. Somebody shows up in a van with a bunch of test gear and replacement parts. They diagnose the problem, swap out a cable or board, and the business keeps humming.
Add IoT and fancy telemetry. Monitor sensors and run ML to do predictive analytics. It still requires someone in a cargo van with a bunch of spare parts.
If you put an installation in a remote location, under water, or in space, there are benefits. But you just multiplied the cost of basic maintenance. You can build in redundancies, but a small glitch at a single-point-of-failure could cascade to a catastrophe.
Microsoft and Google tried putting data-centers under water, and after years of trials they nixed it because of these issues. Multiply that for something far out in space.
Only alternative is to build failure into the business model. Something breaks, you just leave it to rot (or burn in the atmosphere) and send another container out. Maybe the finance math works out, but it sure is wasteful.
tl:dr: SciFi is fantasy. A clogged toilet is reality.
Methinks the reason techbros want to put compute in space is to have it physically away from people. Normal data centers are right there to be smashed and torched in the event of widespread societal unrest. In space, it's unreachable.
The amount of energy required to send a human up there for maintenance is, literally, astronomical. Can a datacenter run at a profit with zero humans and up in space? Your guess is as good as mind.
Data processing in space can be useful, albeit not in the way that Musk seems to sell it. I spoke with some people from a space startup (edgx) at a jobfair in Belgium half a year ago.
Their thing was that they did some data processing in space to cut down on bandwidth (so this is specifically for applications where satellites are a necessary component). I also remember specifically asking them about cosmic rays and they told me it's definitely manageable because their computers fly in LEO. (They already have a computer in space)
Yes because it's a bubble, none of them make real money so it's a waste to actually build them, but you have to fake it to scam investors. On earth the investors will look at land purchases and permits and power contracts and construction, so you have to spend millions of dollars to fake it. You can trade stock back and forth with Nvidia and "order" infinite chips that will never exist but you can't do that for electricians.
In space you just have to already own a vanity project rocket company. You fill out a form with the FCC and get a theoretical permit to launch "up to" 100 quadrillion datacenters into space, and then just point to a windowless factory that might theoretically have the capacity to make datacenter satellites but you can't let them see inside because you have classified defense contracts. Who knows might be a trillion dollars in shareholder value in there, you can't prove there isn't.
It's an idea used to justify merging xAI (burning through cash) and SpaceX (more profitable*). It's not remotely practical due to heat dissipation, general complexity, additional cost, and complete inability to service broken components.
My understanding is that computers in space need to have redundancies in the hardware design because bit flips are a lot more common. Radiation and whatnot. Regardless, you can't just shove some parts in a case and expect it to work. At least not for very long.
Can any kind strangers here share some reads (or anything) on heat dissipation in space? My understanding is that there is virtually no medium to carry away all that energy, but I remember reading something about satellite and space stations actually cool off and there being some sort of a solution to heat. I just wanna learn more.
It's about being outside of any company's jurisdiction. Same thing for the floating datacenters at sea. Can't tax it and regulate it if it's not in any legal country.
If you find somebody willing to pay some 1000 times more for a computer because it's rotating around the sky, then yes, it's profitable. Those "1000" are probably an understatement, but there exists some number that works.
There's probably a single-digit-billions market there.
For comparison, SpaceX's space business is a 2-digits-billions business, and they claim they have a 2-digits-trillion market to grow into.
There's a lot of misinformation circulating about how it's impossible to cool things effectively in space, but it's really not a complicated or new field. I think people mostly just want it to not work and stop there.
top 50 comments