But rewind to 1990, in our time line. Until then, Microsoft were the junior partner in the IBM/Microsoft alliance. They'd promised to deliver a superior OS to MS-DOS, with a windowing system on top; the first couple of iterations of OS/2 were dismal, as were Windows 1 and Windows 2, but the plan was to collaborate on OS/2 2.x and then ship a next generation product, OS/2 3.0, aka "NT".
As it happened, Microsoft gambled on NT while IBM stuck with OS/2, thinking that they still owned the market.
But if we back up to 1987, before OS/2 1.0 shipped, the UNIX-on-PC codebase was clearly a better platform for a multi-tasking OS, and MS-DOS backward compatability could be supplied via VP-IX (if I'm remembering SCO's DOS-box's name properly). If we posit that Microsoft retained Xenix and decided to develop UNIX further, the only question is whether Bill would have coughed for an unlimited AT&T license, just as Sun did. And I think -- given MS's history of buying up the IP they needed in order to hit a marketing requirement (and remembering that I'm an outsider, so this opinion is worth exactly what you paid for it :) -- they might well have done so.
(This was before the dark days in the late 90s when, having been spanked out of the desktop market, the board of SCO sold their IP and the company name to a linux startup from Utah called Caldera, who renamed themselves the SCO Group and went into the IP troll business. Thus indelibly blackening one of the names on my resume.)
If M$ hadn't sold Xenix to SCO and focussed on OS/2 it's possible they'd have moved wholesale into UNIX circa 1988 with the advent of the ia32 architecture. And the whole history of desktop OS development through the late 80s and 90s would have been unrecognizable ...
(Somewhere I have some old PC magazines, circa 1982-84, with advertisements by Microsoft promoting "UNIX -- the operating system of the future!")