> Everything I like about Space X is about doing things that you can't do on earth, because you can't do them on earth feasibly.

I agree those kinds of things are more exciting, but the other angle is SpaceX needs an excuse to do lots of Starship launches, just like Starlink has done for Falcon 9. If orbital compute can at a minimum be profitable enough to pay for the launches it will help bring the cost of Starship down and the reliability up, allowing SpaceX to do more with Starship.

(Of course if orbital compute merely breaks even then maybe the current valuation isn’t justified)