It would be possible to use hardware virtualization features to emulate the different memory layout and hardware MMIO. I would predict that most of the bottleneck is going to be in GPU emulation and the like, so I don't think it would help that much. The greater control and precision of software emulation for the CPU would make quite a bit of sense.
Are GPUs the bottleneck? I thought the modern approach (for anything above an ~SNES) was to simply rewrite graphics calls to run on the native hardware. It's not like Dolphin can emulate 1080p by fully "correctly" emulating a GPU.
