how soon space datacenters come is very hard to tell but I've come to think that if a series of surmountable engineering problems (mainly: Starship; thin film solar; ideally in-orbit manufacturing) are solved, they can rapidly *displace* land-based compute Elon isn't grifting
