As I said that's more like #3. It's certainly very important to their overall plans particularly in terms of going beyond LEO in an economic manner: Starship's

As I said that's more like #3. It's certainly very important to their overall plans particularly in terms of going beyond LEO in an economic manner: Starship's design requires refueling in-orbit for that, and expended tankers wouldn't make the numbers work. They'd be enough for an emergency or single critical trip, so for example reentry is actually not a blocker for SpaceX's portion of the planned NASA HLS program (because Lunar Starship is only for the moon, it's not coming back to Earth, and while it might destroy any profit margin SpaceX could fuel it up with expended tankers if they absolutely had to). But it's not really MVP because a lot of the raw cost/limiting factors of cargo Starship will be in Super Heavy where the vast majority of the engines are. The numbers people have been playing with look pretty comfortably like expended SS+recovered Super Heavy might well end up cheaper in terms of $/kg then F9, and at any rate the raw increase in payload volume and weight means it can do stuff F9 just can't at all. SpaceX can self-bootstrap here and accept the risk thanks to Starlink, so once they can get SS reliably to orbit and recover SH, that's probably enough to make testing at least somewhat pay for itself while they work on recovering SS. Of course, maybe they nail it on the first reentry attempt! But being able to make money doing launch and have recovery be a bonus while the kinks get worked out is exactly what they did with F9 and I think would be a pretty good place to be in.