And yet, you can take a solar cell out to Pluto where the mean temperature is 40K rather than 300K and theoretically get far higher efficiency than 95%. So 95% isn't really a fundamental limit.

Yep, I'm assuming the solar cells are on Earth. Seemed like a good assumption to make.

If you just send them into space with good enough radiators, you'll be able to get 3 9's of efficiency (mid way to 4). But you don't get to reuse the heat waste.

You may be interested on that:

https://en.wikipedia.org/wiki/Matrioshka_brain