Porting AmigaOS to x86 architecture is a waste of time, not only would it be competing with Windows, it will also compete with BeOS, Linux etc.
I've said it before, but what Amiga needs is it's own brand identity to set it apart with a genuine reason to chose Amiga over Mac or x86. Twenty years ago the A1000 had clear hardware superiority over the competition, hence the selling point. Even this wasn't enough though, it required several killer applications before it would become a viable alternative.
As we've seen with Windows, aggressive marketing beats superior architecture every time, but Amiga is fighting for survival with comparatively poor hardware and a marketing budget that wouldn't pay for a round of drinks.
Simply put, we need something unique that no other platform has got. It might be some absolutely stunning game that wins rave reviews, it might be an add on / application that appeals to artists / musicians but just selling a box that runs legacy Amiga software at a premium price is suicide. Likewise, recompling Amiga to run on an x86 box will offer no advantage either.
That isn't to say that Amiga shouldn't adopt Athlon or similar as it's core CPU, as development investment is clearly in excess of that set aside for PPC. Programming issues aside, a top end Athlon based Amiga would be cheaper to build than it's PPC counterpart.
