I think it's going to come down to the business case.  If Windows ARM isn't a viable option, either because it stays an OEM license, or because the x86 emulation won't run inside Fusion, then there's a stronger need for Apple to do something.  But if Windows ARM virtualizes x86 well, and runs in fusion, and has a retail license option, then I think that would probably cover 90-95% of the use cases.  I doubt that supporting virtualizing of older OSX versions on ARM hardware much factors in to the equation.

I expect to see ARM on ARM in weeks, rather than months.  Then it's up to Microsoft to give us an ARM windows option.