Does anyone have a good understanding of why PyPy did not supplant CPython? I'm not familiar with it, but from reading their website, they claim it's better at pretty much everything. So the question naturally arises, "if you're so smart, why aren't you rich"?

Good question. One reason is PyPy didn't support NumPy until last year.

Main reason is for majority of the workloads it didn’t improve performance significantly.

In some cases cpython was much faster. Didn’t offer much incentive to move to pypy JIT from cpython JIT.

CPython implements a simple bytecode interpreter. It does not have a JIT.

Lagging Python 3 implementation early on didn't help.

https://news.ycombinator.com/item?id=11541744

too many libraries don't work with pypy. The last time I checked, psycopg2 doesn't work. Completely a deal breaker for me.

Just out of idle curiosity: Why don't they work?

PyPy doesn't completely implement Python C API (which anyway is specified only by the CPython implementation, not in any formal manner).

There's psycopg-cffi which does the job

Because CPython is the reference implementation.