The difference here is that the previous ISA changes have been accompanied with massive benefits. The 68k to ppc transition came with ppc being able to emulate 68k code faster than any 68k you could buy (in an interpreter, no JIT!!!), and the native ppc code was just ungodly fast comparitively. Ppc to x86 was less dramatic on raw perf, but was accompanied by a huge gain wrt perf per watt, and you could still emulate ppc on Intel about as well as you could run on a native ppc (with a JIT this time).
But now, Moore's law is in the process of hitting a brick wall. Those big wins for free just aren't there. And if there's a x86 to arm JIT, it's either going to be single core, or the arm cores are going to have to have a stronger x86 like memory model, thus negating a huge part of their architectural win.
IMO, Apple's going to make their own x86 chips because the base x86_64 patents are going to expire real soon now, and they have enough patents from the crazy amount of fabless acquisitions to potentially cross license the rest.
Processors are commodities now, and that's really going to shake things up in crazy ways.