[GMoromisato]

I'm not sure I get it. Obviously, computers can work in space--a Starlink satellite is basically a computer with a radio attached. Satellites use radiators for cooling without violating any laws of physics.

I assume you think that SpaceX will never be able to build/deploy a radiator big enough? But that's not a physics/thermodynamics question, that's an engineering question. And I think SpaceX has some pretty good engineers.

Help me out and tell me how you can be so sure it will never work.

[jquery]

You could also launch people from LA to NYC via rockets instead of using airplanes without violating any laws of physics. But we don’t, because we have airplanes already. What problem are data centers in space solving? Cooling is probably the hardest issue to deal with for data centers, so why would you give up convective cooling you get for free on Earth?

I don’t know why you’re jumping from starlink to data centers in space. The utility of satellite internet has been known for decades before Starlink came around.

[9x39]

>What problem are data centers in space solving?

A hedge that as it gets harder to build datacenters in communities over water, power, noise, space, tax, etc reasons that space is a new frontier for them. Consider that solar power works better in space.

Also, satellites may be able to do processing of data in-orbit.

>I don’t know why you’re jumping from starlink to data centers in space.

You just mentioned it - cooling. Consider Starlink a POC of radiating computer heat in space. A datacenter would need a scaled-up version, but it is not impossible, although it could be impractical if the cost of compute doesn't rise enough.

[GMoromisato]

That's a different argument.

You agree that space data centers are physically possible, but you just don't think they will be economical (i.e., cheaper than terrestrial data centers). Is that right?

I don't know if they will be economical. But that will depend on a whole bunch of questions that nobody knows the answer to: How will the demand for AI grow? How much will opposition to terrestrial data centers increase the price? How cheaply can SpaceX launch mass? How cheaply can SpaceX build data center satelliters?

Maybe you know all those answers. If so, I envy your stock portfolio.

[microgpt]

They won't be economical for many reasons - one of them is cooling. Putting several kilowatts of radiators on a Starlink satellite is a justified cost because the Starlink satellite must be in space. Putting a hundred megawatts of radiators on your AI DC in space is ridiculous because that's too big of a radiator (how do you keep it structurally sound?), air cooling on the ground is a million times better and there's no good reason for it to be in space.

[rlt]

> won’t be

Again, “What I find silly is the certainty that critics have that SpaceX will fail”.

I doubt you’ve done enough investigation to conclude that. Watching a couple a YouTube videos about how data centers in space are dumb doesn’t count.

For one, it’s not one big hundred megawatt-scale datacenter, but many smaller rack-sized satellites.

Two, there are many trade offs SpaceX can make to reduce the overall cost. The radiated power is proportional to the 4th power of the temperature, so anything they can do to increase the temperature drastically reduces the size of the radiators, including custom silicon, heat pumps, etc. There’s also novel radiator technology like liquid droplet radiators that could be worth developing.

Third, there are reasons to put it in space: solar panel efficiency, no cost for land, less regulation, no NIMBYs, etc.

[microgpt]

Materials other than silicon would be an interesting idea, but I doubt even SpaceX is that crazy.

Not only does cooling increase with 4th power of temperature as you said, but power is abundant, so it may actually be practical to use more power to run chips of a different material at a higher temperature. That's a great idea, if it can actually be pulled off.

There's nothing inherent about electronics that makes it stop working around 120 degrees - that's just the practical limit of silicon. Other materials can withstand higher temps but may not be as practical in other respects like bandgaps.