An interesting note is that Autodesk started out with Autocad, which they coded in lisp (eventually transformed into autolisp.) They took the MS model to software development and proceeded to make bigger and more complex versions with increasing quantities of bugs to the extent that multiple crashes per day were the norm. But they owned 80% of the world cad market, they became the 2nd largest software company in the world.
Revit was developed by a small company with an incredibly sound software architecture with a database at the core(without the massive bandaids that had piled up in Autocad and are still there today.) and developed a devoted cult following. Autodesk fought back with "Architectural Desktop", which I bought (before I knew about Revit). Their version: Cut a section through a building, wait 30 minutes. Revit, Cut a section through a building - instantly. Autodesk got very scared, then bought Revit. They saw the writing on the wall and undersood that Revit would be completely replacing Autocad in the next few years, which it is doing now.
So if MS can buy Google, they'll be all set :)
As much as I am frustrated with having to use XP (My Autodesk BIM cad system only runs on it), I wouldn't underestimate MS's ability to bounce back from bad directions/decisions like they have in the past.
If Gates radically changes his vision and business model to the new web world and gets Balmer to agree, they could dominate again. But can a company that has written extensively buggy software in the least lean way transform at this late date?
If it were anyone else other than Gates at the top of the company, I would say no. But does Gates have the energy this late in his career to make such a dramatic change in his business? Maybe not.