Hacker News Reader

Yeah, I have no doubt that it's possible to put GPU servers in space. The Starlink satellites are servers in space. The question is whether it's even remotely profitable to do this for AI training/inference servers.

View on HN

The same way that xAI has compute on the ground and there's such demand that the compute exists so it will be sold. I think SpaceX can conceivably sell the space compute easily. This goes with the business strategy of Elon which is just sell services to himself. He sells rocket launches to himself with the excuse of creating a space datacenter. The same way he sells cybertrucks to spacex because the public won't buy them. I'm pretty confident it's definitely cheaper to put the compute in Texas, but might be viable enough to sell himself rocket launches and make his company worth more to leverage for more ventures. I'm not a business person I don't really get how it's all so valuable if the only demand is itself.

Training, no. Latency is too high.

Inference the latency becomes trivial.

Other things, I suspect latency is too high again.

They're going for an orbital height of 600–800 km, or about 2-3 light-milliseconds from Earth's surface. A lot of people think of satellites as being necessarily high-latency because of intuition drawn from satellites in geosynchronous orbit – but those are about 50x farther away from earth and have correspondingly longer ping times.

They don't go directly to the satellite but through ground stations and then bounce around on lasers between satellites. The latency is going to be worse than fiber

Reply on news.ycombinator.com