The entire theory has a few _massive_ holes in it.
Operating systems aren't particularly useful on their own. You need applications. Windows has a massive ecosystem of those. Millions of them. And approximately zero of those will run on your nice new ARM notebook. They are all compiled for the x86 micro-architecture and won't run on ARM....not even with compatibility mode and fairy dust.
So essentially MS will have to convince people to buy an overpriced OS for which no software exists. It gets better: They'll have to compete against 2 tried and tested solutions that essentially do the same thing minus the problems: Windows on x86 and Android/Linux on ARM.
Another factor that these fine analysts keep getting wrong is ARM's price advantage: Its an illusion. Intel is king of the hill and charges a premium because they *can*. If Intel feels there is danger in their core market though, they'll drop the premium (temporarily). Then they a price advantage (or parity), a performance advantage and plain x86 inertia on their side. Good luck competing against that.