I assume this is for dedicated unshared bandwidth with no cap since it's for ISP. If that's the case, if it uses the average of one satellite that's 3+ million SpaceX needs to gather to break even (not counting the cost of operations or equipment itself). As far as cost competitiveness installing and maintaining an undersea cable is a worse deal (though preferable if you can afford it). Overall the prices don't shout "ridiculous!" to me by either measure. It actually seems pretty cost competitive and probably not full of 1000% margin or something.

Via another comparison it's a 10x increase vs 1 TB of priority data @40 mbps (250/m) but that's actually a good value considering this is @1000 mbps with no cap. Shooting 10 Gbps (10x cost) to the island, for example, is likely to cost a bit more than it would have to have the island's population all by Starlink individually but they'll get significantly better service as a result.

This is something that's only viable for pocket areas in remote locations though, particularly islands. Dense areas or plain rural areas fit either normal connectivity or traditional Starlink better.