QEMU is a great project, but to begin with flexibility was not a primary design goal. Majority of QEMU targets are PC-like hardware and older consoles have unique or alien architecture. To emulate it using QEMU you'll have to patch it internals heavily. It's also written in C so you can't just integrate a lot of fancy libraries into it.
QEMU is also important infrastructure project playing crucial part in *nix virtualization stack. Which means getting any code upstreamed is hard. Emulation projects primary goal is to make games run both accurately and as fast as possible which means developers will use as many shortcuts as possible.
QEMU is a massive code base too. Getting into it will take a lot of time and it could be easier to maintain single-target emulator instead since it's easier for new contributors to start working on it. You'll have to fork QEMU anyway and won't be able to push your changes upstream then what the point to use it?