I want some of what you've been smoking for Christmas.
I'm no fan of Microsoft, and I've been running a Linux desktop since the late 90s, but I think your plan is stupid. Windows is not way behind in scalability, other than on extremely large servers (and the market share of servers with hundreds of cores is not large enough to change Windows like that) If you think it, show your evidence. Benchmarks on sites that run software that is available on both Windows, Mac and Linux show otherwise - the underlying OS has very little influence on performance except for stuff where the drivers are heavily used (primarily graphics, where open source drivers are generally behind Windows)
Microsoft is not going to gain any customers by doing this, so why invest a billion dollars or whatever it would cost to do it? There's no benefit to Microsoft, no benefit to its customers, and risk that it makes the world easier for Linux or BSD or whatever they chose for their new kernel.