The only problem that "data centers in space" solves is the problem of trying to scale a rocket company where the potential demand for rocket launches is simply not that big.
There are a lot of problems that can be solved by creating 20 other, much bigger, problems.
It remains to be seen how efficiently those engineering problems can be solved, but none of the commentators here have any real insight about that (myself included). My money is on the ingenuity of the engineers that are working these problems.
So the money we are saving is commodity solar, commodity steel and community battery (or any other form of power).
And instead of those things we add, insane engineering complexity, insane complex heating, insanely complex maintenance or standing backup, insanely expensive transportation and insanely complex operations and insanely complex communications?