Besides, I don't see any person/people being able to understand such a complex feat of engineering (and legacy) sufficiently to maintain it for years to come.
I would assume that there is enough interest in maintaining Windows - especially commercial interest - that finding and funding a sufficiently large and knowledgeable team is actually not unrealistic. It's just unrealistic to assume that the source code will ever be published willingly by Microsoft as long as any part of it is a direct ancestor to a product they still sell.
As for the commercial interest part, just thinking about how video games' source codes are so rarely published saddens me a great deal. Even after the commercial interest is lost and the IP is no longer in use, companies don't really think much about preserving the source code.
https://www.theregister.co.uk/2019/07/03/reactos_windows_res...
What about open sourcing, say, windows 95? That seems more achievable, and probably more helpful.
Compared to that, Windows 9x was less kludge and more a straightforward implementation of Win32, though of course lacking modern features.
The Windows NT4 source code leaked a while ago and can be found all over the internet. Of course that doesn't really help ReactOS because they officially can't use it, but that's probably the closest we will get to MS open sourcing Windows.