It is just that we see many of these new concepts and not all come to fruition. Even if this one does, it is likely to be some way down the track, I can’t see MS or Intel abandoning existing revenue streams.
So this I feel will be an alternative stream until it builds up a point where cost V benefits reach a point where existing technology cost V benefits are matched or bettered. If you want this, you not only have to change hardware, OS and if supported software needs to be ported to ARM, then it is going to be very costly.
As a Mac user I can say that switching processor architectures does not have to be a painful process. Apple has done it twice (Motorola 68k > PowerPC > Intel) and it was a smooth process for users and developers. As long as Microsoft makes some sort of compatibility layer to aid in the switch it should not be a problem.
For example, I still have applications that are not written for Intel (example lots of my older documents are still in AppleWorks which is no longer developed) and it runs fine with the Rosetta translation. Not every developer will see a need to port to ARM initially, so to have some kind of translation software for existing x86 software to run will be key to its success/failure.
Windows on ARM should be a great way for low income users to get a PC or at least the threat of it should get Intel to come down on the high cost of its Atom chips.