Well, there's Wine and ReactOS as two major efforts to reimplement Win32 user space and the operating system as a whole. These are big team efforts, certainly, but they manifest a very detailed understanding of Windows, mostly obtained through reverse engineering.

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.