Free advice: The economics don't work out.
post
Getting rid of the heat is going to be an issue for that... along with the massive pollution from the many launches required to get this in orbit.
The heat will just dissipate in the air, and they can launch it at night when it's colder. Science!
/s in case, there are a few mouth breathers out today
They could build them so that they stay in perpetual dawn or dusk. One edge with the solar panels in the su, the other edge with the cooling fins in the night’s cool breeze.
Geostationary orbit is far higher than low earth orbit and I would assume following earths twilight zone would not be much better. I do not see why you would either, with reaction wheels you could orient the satellites towards the sun regardless of the relative position of the earth, with the caveat that earth may block the sun which is hard to avoid entirely anyways.
Also, there is not that much cool breeze in space, famously known for not having vast amounts of air (still have IR-radiation to help though).
Edit: Probably ate the onion, didn't I?
Oh! Oh! Attach it to a meteorite! Almost infinite heatsink in space!
I'm pretty sure they're aware of the need for radiators. They've probably designed satellites before.
Nobody thinks they're incapable of working this out; we think theyre deliberately advertising something dumb that lay people won't necessarily understand is dumb. Replying that they have smart engineers is stupid because no-one denied it - we just don't think they used those engineers to come up with the idea.
yes but they're not trying to dissipate megawatts usually
Again, I'm pretty sure they're aware that you need bigger radiators when you're using more energy. This is space engineering 101.
literal kilometers of panels and radiators. No. It won't happen
If only they'd hired you, they would have known.
well any actual engineer who isn't trying to sell them will readily tell you that a datacenter in space is a very bad idea.
They'd better not try to sell them to anyone who has access to an engineer, then. Just a single engineer will bring the whole scheme crashing down.
for starters, at the loads they're running at, they have literally hundreds of gpu failures a day. How do you propose doing that in space?
Include spares.
I hope they're reading this thread and taking notes, they probably didn't think of that.
and the infrastructure and robotics to replace them, of course.
Assuming 200 nvidia H100 failures a day (conservativo, reality is worse) that's an extra ~340kg of weight you'd need to launch per day. Which is an extra 120 tons yearly.
So, one Starship launch per year. Doesn't sound like a problem.
at least two, you can't stuff a rocket full of just gpus, you need something to actually dock and deliver the payload in space. So you need to launch at least 2 rockets (in a non-reusable configuration, so you need to pay for the whole rocket and the launch) to ship a bunch of gpus that are, at best, only 10% as fast as usual.
and forget about running 4nm chips in space. shit has to be radiation hardened, which means bigger process nodes and higher energy cost, and lower speed
Another thing they probably didn't think of. Nobody's run chips in space before.
they did think of it. lots of people have. I just mentioned what was required. Rad hardened processors are usually 10 to 20 times slower than what we have on the surface
So don't use rad-hardened processors, put them in a radiation vault instead. Those become more mass-efficient the more hardware you're putting inside them.
Really, I assure you the people proposing these things have put more thought and expertise into it than a bunch of random Fediverse commentators.
Edit: don't feed the troll.
Me too. I'll even make them full AI.
Please send me $2 billion by Tuesday. My salary as yetAnotherUser CEO & CTO is a modest 20 million/year. Results are expected to appear by 2030.
Hey Gemini, make me a business plan, a marketing site and some presentations with fancy graphs.
Hey hey, I'll make them on Mars! Send the cash to me instead!
I'll do it right here on earth for only 75% of the cost. Think of how much you'll save versus mars.
Imagine getting sent up because a pos rj45 cable got hit by a micrometeorite
Because they swore the server was powered on...
Back when the capacitor plague hit I had to manage locating & replacing over 500 motherboards in the datacenter of my then-employer. Imagine if a hardware glitch like that happened in one of these.
I had some of these bad capacitors blow in my video card, back in the day. I was extremely proud of myself for managing to order some replacement capacitors and soldering them in myself.
The most impressive part might be that I ordered the right items. I knew nothing about electronics repair at the time, I just wanted to be able to play World of Warcraft again.
Isn’t it incredibly difficult to shed heat in space since the only real way to move heat is radiation?
In the (fiction) novel Artemis by Andy Weir, which takes place in a city on the moon, they have a heat management system that seemed pretty cool. They convert heat to light, and radiate the light out into space. Not sure how feasible/scalable that is, but I thought the concept was cool.
If they can turn it into light why not turn it into energy to be used?
Doesn’t that already happen with infrared radiation?
That's already how they do it. They pass heat onto radiators which radiate away the excess heat in infrared. It should be noted however that this is far less efficient than it is in an atmosphere.
There are some various ways. Radiators can be large and thin, and as long as the heat-sensitive part of the thing is kept cool it doesn't really matter how hot the rest of it gets.
In before someone wants to reboot a server and the hypervisor is unresponsive.
Data centers in the ocean didn't work out. This doesn't seem smarter
I didn't realize there was a paywall, I use the bypass paywalls clean extension.
Come on now, people can't actually be humoring this fever dream, can they? It's just so fucking stupid...
Isn’t this how Skynet started?
News headline in the future: "Orbital Data Center Crashes Into Los Angeles in a Mass Casualty Event"
Even if this was an economically sound proposal, the next X45 magnitude solar flare might be a nasty surprise for reliability metrics...
Edit: at some point, this would also likely contribute to Kessler Syndrome, but at least we'd have chat bots.
Do you want furries in space??
Bcs that is how you get furries in space.
FUCK OFF Thank you
> People complain about the environmental footprint of data centers.
> Companies attempt to move the data centers outside the environment.
> People complain even harder.
What do you want?
What do you want?
Realistic, sustainable progress instead of endless, phony, destructive grifts.
Kinda obvious.

top 74 comments