What can you even do with 20 qubits?
This isn't the right way to think about this. No amount of qubits would allow you to outperform a classical computer on "most" real world problems. That is not how quantum computers work.
However you are right - anything quantum right now is proof of concept for the engineering problems, not being problem. (If you are being generous. If you are being less generous its about creating PR to get people's money)