Intel Corp (INTC)

[Unkwn]

They would be running a JIT to translate the code.

That requires your code to be present as IL in the first place. I doubt this is the case for a large part of existing Windows applications.

For all those applications, including the driver layer as mentioned already, it would require a form of emulation which is generally slow. In addition, for low level stuff done in drivers, things can't just simply be emulated. This needs adaption of the existing assembler code by hand. I guess this solution will face the same driver and compatibility issues Windows RT had. This all boils down to the question: Will Qualcomm achieve a performance that is superior to competing Intel chips (respecting TDP) to offset the compatibility issues and performance penalty? I highly doubt it at least for the coming decade. I fear much more Qualcomm's server ambitions. Server is more about platform architecture, something Qualcomm is really good at.

The reasoning for Microsoft seems clear to me. For them, ARM is the dominating ecosystem (it is in terms of units) and they want to get a grip on it somehow. Wintel is dead for sure, they are just sitting in the same boat together.

In the meantime: Samsung announced to fab Qualcomm chips at the beginning of next year on its (so called) 10nm process. They claim about 30% density improvement. Since these chips are expected to be used in Smasung's upcoming Galaxy S8, there is not much doubt about the schedule. If this holds remotely true, Intel will have lost its process lead by beginning of next year. Anyone thinks TSMC will beat this? TCE?

In the meantime, I think Intel is up for some PC recovery. There are finally some really nice high quality laptops on the market and there is some vacuum to be filled, since people refuse to use tablets and smartphones are just too limited for. I think, Intel is a sell by then. The company just can't get over its corporate culture issue it seems.