Starship would be really good even if they can't make it reusable, but still hit the early-user price tags and capacity to LEO they're talking about for expendable-mode launches, even if they then stop improving.

But unless there's demand for launching a lot more stuff, that's not going to help them make a lot of money.

I can buy into a vision of several competing Starlink-like constellations. Can't buy into one about space-based data centres. And we need a lot more demonstrations of space colony precursor tech before I'm going to even have a strong opinion about the economics of space colonies*, either for or against: they're technically feasible, but so too is one on the sea floor, or the middle of Antarctica; and there's people who already build homes for themselves in the middle of the Sahara and at high altitude on mountains, but these aren't *huge* money-makers, just standard boring ones.

* they are cool though; if Musk had actually delivered on the Mars flights that were supposed to have already arrived two years ago, and also not turned out to be who he turned out to be, I'd have considered being an early Martian: https://techcrunch.com/2020/12/01/elon-musk-says-spacex-will...