windows succeeded on the desktop by providing an ecosystem where participants could build on/with each other. and yet windows interfaces were proprietary and closed enough that no one could ever really clone the basic Redmond-supplied parts.
but now, the world is changed: the ecosystem is the net/web. now there are actual standards (open and free, not closed and proprietary like windows). it's hard to see how Microsoft has any real traction online, since the whole point of any open/free interface is that it's possible to substitute either side.
WIndows on ARM is a strong sign of competetion ahead for the Intel and AMD in the PC and Tablets market. With many vendors releasing better computing processors having inbuilot graphics controller will be a general trend for the future.
An OS is a platform for running apps. Where are the apps that run on Windows-ARM? On the Android platform, of course. Okay, so MS is looking to displace established players using a technically less energy efficient OS.
What is the point of running Windows on ARM? Think about why people run windows at all - there are basically three reasons.
One is familiarity - but by the time Windows on ARM comes out, most potential users will be familiar with Android.
Two is "it's what comes with the machine" - the same or similar models will also be available with Android or other Linux.
Three is application compatibility "I need to run this windows program". Without x86, this point is void.
So what's left? People who want to run IE and MS Office, and no other programs, and want the slightly lower costs and higher battery life of running on ARM than on an x86 (a future version).
Microsoft will do best (and I believe they know) to optomize for Fusion APUs; APU's provide the best platform by far to show off Microsoft programming and innovation (yes I said the 'I' word relative to Microsoft). The 'x86'(x86+GPU) platform just got stronger with AMD's Fusion.
Huang tends to embellish a bit , from past interviews- . Question is, how long will it take to learn and incorporate their own CPU; this isnt a leggo project- or could it be a last grasp. The other bit about Microsoft/ARM O/S is getting way to much attention- firstly ARM designs are specilized, and do not lend for a one size fits all O/S, second- the ARM Consortium group, who asked for a WindowARM OS, Qualcom, TI, I forget the other , are creating their own branch, which will probably feel comfortable with X86 coded programs running on their hardware- which could present big design challenges- perhaps towards the middle of 2012,by that time Fusion will have cemented its self, and who knows how far (money is no object) Intel will go to reclaim its position.