It wouldn’t need to be GHI, but I agree, other MS or Intel doing it, I do think it’s unlikely.
The only way I would see Intel doing it is expand the surface area of the battle against ARM based systems.
*** thinking out loud in my ignorace ***
This has a lot more going on with this setup than a typical mCu based system. Correct me if I am wrong, but the Quark is more CPU than mCu. Most of the I/O is done outside of Quark vs the integration of an mCu.
The benefit of the Quark, if it continues be a non-SoC, is that when Intel produces faster chips or more energy efficient or both, as long as they are instruction and pin compatible, you could “just” use that instead.
Plus Intel is open to modification of the cpu and doing fab work on this. So if a company wanted to build a product around the Quark and thought "if only it could do this in h/w as well … ", now they can have that, for a price obviously.