> No matter how cheap they manage to make SpaceX launches realistically, there's really no situation that a space datacenter makes any sense compared to putting datacenters in, for example, Antarctica.

Solar energy is going to expensive in Antarctica.

We can imagine a situation where the server hardware becomes so cheap that the energy cost dominates. In that case, sticking the things in space could make sense, particularly if extremely low mass space PV (just a few microns thick) can be made to work and also work cheaply.

You still would have to deal with the fact that you would almost certainly need to budget at least 2x the regular amount of hardware to deal with the fact that you can’t do stuff like “replace failed power supply” or “replace failed hard drive” without launching an astronaut into space, so you would need to have an abundant amount of resiliency to overcome that. You know, fault tolerance. Something you can’t handwave away for a data center.

I am harping on this point because You can’t just say “but in future computes won’t need maintenance” because at that point you’re just engaging in speculative fiction that you have no way of knowing. I could say “in the future we’ll have cold fusion” and maybe that’s true but I have no way of knowing.

So given that you would need 2x the power that you’d need on earth. Compared to just putting a shitload of solar panels in a desert where non-astronauts can easily access it I don’t see the point.

And of course there’s the nuclear-power-plant sized elephant in the room; if power is the constraint it would still almost certainly be more economical to have a nuclear reactor than trying to get a data center in space.