IMHO, if you need to emulate the 68k programming model in the first place, the only point left is speed. There's no point in starting with a comparatively slow ColdFire in the first place. The bus is hard to adapt as well, so no advantage gained there either.
From the speed POV, you're left with PPC and x86 - with x86 being highly available in all speed grades and form factors. Rather than designing a new accelerator board, it'd be much easier (well, somewhat) to redesign the chipset (when you're hooked to it) since this is already worked on. Put all this as an FPGA on a PCI(x) board, use an off-the-shelf x86 mobo and you're set.
Actually there isn't really too much point left in dragging the chipset behind either - if you drop that as well, you're left with the OS. Port it to the platform of your choice - et voilá - that's where AROS comes into play.
It all depends on what you really want to achieve:
- save your old hardware from dying (expensive)
- modernize your system while keeping old software (=> emulation)
- regain a good price/performance relation (=> OS porting)
Everything else is :horse:
