

If and only if someone is insane enough to develop off-planet manufacturing with the bulk of the raw materials originating from somewhere in deep space, e.g. asteroid mining, putting data centers in space might be useful for problems that demand intensive compute and can work with extreme latency.
Then again that’s like saying inventing the airplane would have been a good strategy for Neanderthals to find better firewood.





He’s stupid, but in this instance I think he’s just bullshitting for spacex’s sake. Even he is capable of putting together the fact that the costs required to build (not to mention service) an orbital data center far outweigh the costs of swapping out the genset for the same exact solar panels you would have otherwise put on the rocket. Maybe you need more panels to make up for atmospheric transmission and the day/night cycle but he (or let’s be real, grok) has ready access to the cost of a single starship launch and a calculator.