Is there literally any justification for this or have they just completely lost their minds
Well, there's "Imagine how cool that would be, bro".
And they are kinda right, if you imagine it a fair bit cooler than it would actually be…, bro.
Other than that, you'd have to pretty much disregard anything learned from this. And you'd have to disregard the entire issue of getting rid of heat, powering a power hungry system that's in earths shadow half the time, maintaining a server farm that's constantly 500-1500km away from any technician and also in a vacuum, along with managing a solar flares, 500ms of latency and bits of sand traveling at 10km/s hitting your servers - oh and your boss being on drugs half the time.
Doesn't sound very "cool" to cook your servers with waste heat in the vacuum of space lmao
It wouldn't have to be in the earth shadow half the time. A highly eccentric orbit wouldn't have that issue. It would have the issue is constantly changing latency, but maybe at the slower speeds AI works at that wouldn't be a concern?
some investors think it sounds cool so you get a short-term profit before the whole idea falls apart. or something like that.
That makes absolutely no sense.
Data centers are "centers", because they require vast amounts of resources to run appropriately. Resources we don't have in the vacuum of space. You can't even dissipate heat, which is the primary thing data centers are screwing people the world over on right now.
You can’t even dissipate heat
of course you can, by thermal radiation. Power by solar panels which can generate a lot without atmosphere filtering sun energy.
The ISS has a lot of big solar panels. The other big panels they have are thermal radiators.
They have to have quite large thermal radiators because it's very inefficient. The ISS has people and a very small amount of computing power.
Data centers generate several orders of magnitude more heat. You would need several orders of magnitude more thermal radiators than you would solar panels. The bigger you make the data center, which is important for density since you're introducing a lot of lag due to the speed of light, the less room you have to put thermal radiators or solar panels.
Then you need to work out how to get spare servers, and/or server parts up and down from the Data Center. All of these things are consumables, and all of them have significantly more wear and tear outside of the Earth's atmosphere.
It is possible. It is not efficient or sensible. It sounds cool, it doesn't require buying land, and there aren't currently international agreements about doing dumb stuff in space in the same way there are for doing dumb stuff in the ocean.
More so these are not giant data centers they are just satellites acting as a data center. So perhaps larger than a typical communication satellite but nothing like a giant warehouse that we would have on the ground.
Isn't one of the big complaints against data centers is the massive demand on water used for cooling? How are they going to cool these in space which is notoriously hard to cool things down in?
Tech bros aren't actually smart people, they're sociopaths who know how to manipulate people. They have no actual value.
This is also 100% true for MBA's
They would use radiator panels which automatically swivel so they’re edge-on to the sun.
I think the bigger problems are;
If someone knows the specifics on them I would be happy to know, but I feel like there is a lot more heat generated in a data center, of any useable size, than could feasibly be cooled using radiators.
I would like to add a bullet point to your issues since I'm not sure solar would be good enough to sustain such an orbital. Once again I am only surmising and would like to hear if I am way off base.
they will need to be constantly replaced.
This is the funniest part. Can you imagine completely rebuilding the average datacenter every half-decade?
Fraction of damage and cost of earth based datacentres.
We've been putting computers and satellites in space for decades.
That's what shielding is for. We have probes still working after 48 years in space.
so place them in higher orbit.
Well using water for cooling is a way to save money in some regions, but it doesn't even work everywhere, most data centers don't use a lot of water.
When it comes down to it, you can have a datacenter anywhere, including in space, but you do need to keep it cool. Cooling can be harder or easier in different environments, space is probably one of the hardest environments to keep electronics cool in.
I but I guess the most direct answer to your question:
How are they going to cool these in space
They'll do it with radiators, lots of radiators. And they'll do it at 50x the price it would cost on earth. With that in mind, I welcome the space datacenters, build as many as you want. I can't think of any better way for an AI company to drive itself to bankruptcy.
How are they going to cool these in space which is notoriously hard to cool things down in?
https://en.wikipedia.org/wiki/Spacecraft_thermal_control
The problems with datacenters are power consumption and water use. This addresses both problems. Sunny side picks up tons of solar generated electricity, dark side radiates heat.
Radiative IR cooling is still orders of magnitude less efficient than convective cooling. There is precisely zero percent chance that orbital data centers will be less expensive than terrestrial ones.
Does it address water use? I'm aware spacecraft have thermal management systems. But also with the chief complaint of data centers being water use, are these thermal management systems able to dissipate the same energy? You don't need that much water for earth based data centers without an underlying massive amount of heat to get rid of. I'd wager that heat is much higher than what spacecraft typically have to get rid of. It's the actual quantity that I'm questioning.
Death ray from space?
the only person that benefits is the world‘s richest asshole who happens to own a private rocket company.
Ive heard china is rapidly catching up. End of last year they were launching either same number or close to what musk was launching. I thing that why he IPO Ed the company - high time to sell.
Maybe he additionally beefing up the price drumming many launches from space data centers. But as many said seems like those are dumb
Yeah this seems like hype before the IPO. All of the investors who don’t know how any of this technology works aren’t gonna care about anything except having a viable business case that happens to be expensive and involving the keywords AI, data center, space, and rockets.
Once he gets his money he’ll move onto the next shiny object that captures his attention with his ridiculous amount of money. Time to find the next route to more investment money, contracts, and some other shiny IPO wrapped up in the sci-fi dream then only idiots can believe.
The only surprising thing honestly is how many smart people he attracts actually make his businesses somewhat functional.
I would have thought the people heading up Google were at least not as stupid as musk. But here we are...
How about you put their CEOs into orbit?
Well..I guess that TECHNICALLY solves the increased cost of living problem. Though now our taxes will be paying for it...so not much better? Assuming this is even possible to begin with
Project Suncatcher
Space cadets have gone full cartoon villain
Out of reach of the rabble, eh?
Good. The more stuff we can get off this fragile world the better off it will be.
Yes, it's hard. Now at least. But at some point we can source the materials out there instead of having to launch every bit of it and that'll make everything easier in space.
No more manufacturing polluting the air. And once we can source the materials to build, it's a simple matter to source the materials needed for manufacturing as well.
The biggest complaint of EVs is the mining of the parts needed to make it.
Imagine no longer needing to mine the earth for materials.
source the materials out there
That's 5 words and trillions of dollars worth of start-up cost.
If we're so dead-set on flying servers out to LEO so they can be micro-meteor target practice, maybe we could leverage the no-water, solar powered setup we'd have to do there to get better DCs on earth here, first. We'd have gravity and thus convection on our side as we look at cooling, not to mention the vastly better proximity to supply chains during the proof-of-concept phase.
It's really not. Solid state devices don't care one whit about gravity or any lack thereof. And do you have ANY idea how many small things we have today to make our lives better that are all offshoots of the original space race? Velcro. Memory foam. Just to name two of them.
I swear, some people just refuse to see the forest because of all the trees in the way.
There are multiple trillions of dollars worth of materials in the asteroid belt.
Also, solar works best when there's no atmosphere between the emitter and the collector. Not to mention there are orbital vectors available that would keep the collectors in full sunlight permanently.
No massive battery farm needed. No need to go dig up all that material and continue destroying the environment here by doing so.
Why are there so many luddites using technology like the Internet these days???
Velcro. Memory foam.
Seriously? that's the best you could do?
I was just going off memory. And I'm turning 50 soon. Sooo. 🤣
The extremes of space and the engineering of components goes far beyond just gravity. While there is no doubt that new technology would be developed this is not the government doing it. You really think they are going to give this technology away for free like the US government did? Highly unlikely.
It is not about refusing to see the forest because harvesting asteroids, then returning them to our planet for use on the ground is bonkers level aspirations. It would make more sense to keep them in space for manufacturing, but we are very far away from any sort of space building platform that could process raw materials into a finished product.
Energy from the sun is indeed abundant in space, but the costs of sending materials up will remain prohibitive until we no longer need rockets. That type of technology is still very far away although some proposals like a launch rail cannon, space elevator, or ablative laser propulsion could eventually solve this issue.
You also are missing just how fast technology is moving. Current AI centers expect to swap out hardware every three years or so. Who will be up in space swapping out the hardware of possibly millions of satellites. This is another added cost because realistically they will just have to keep launching.
You last comment completely misunderstands who the luddites where and what they wanted. They were against technology taking away livelihood not the technology itself. Your comment completely ignores this reality and paints luddites as something entirely different.
The government didn't do it before either. They contracted private firms to do it, like Boeing and Rocketdyne.....
NASA did do a lot of it directly so no. There were subcontractors of course, but they were usually responsible for building specific things not the research that generally came from public universities and NASA.
SpaceX is not sharing their technology in the same way NASA did. In fact, they aren't even patenting their tech as to keep it secret.
Could someone explain the appeal?
Molotov's don't work in space?
Literally the only reason you would want to do this is to physically move it outside of the reach of the general public.
The ONLY reason I can see for gov't or big-tech to be wanting to be doing that, at the obscene-cost of doing it,
is non-accountability: no oversight, no search-warrants, no nothing, ever.
Permanent non-accountability, permanent exemption-from-rule-of-law.
Which is itself good reason to prevent it from happening, were gov't actually-responsible..
_ /\ _
This is a most excellent place for technology news and articles.
