So one thing you need to understand about MSFT is that they very - like other publicly traded companies - publicly consider lack of growth to be the same as death. In fact, if a divisions numbers are slowing down or starting to stabilize, SEC reports show they will without hesitation axe it in its current incarnation, then re-brand it after a sufficient delay and a little investment from engineering, and tout it something completely different because - generically and overly broadly speaking - it not has a new feature bolted on.
There is a quote that is often bandied about on campus over in Redmond, and its very simply stated as "Re-orgs happen". Everybody who works at MSFT now, or has in the past, knows this.
I suspect based on my personal reading of the links above that this is whats happening; and I'm willing to bet money that a re-org is involved somehow. That in my opinion means that 2 technologies are competing for dollars and once it gets sorted out, the MSFT marketing machine we sometimes love to hate will be right behind the one that wins.
If you REALLY want to make sure something of XNA survives, my personal suggestion - and this is again just my opinion based on hypothetical - is to not let the open source alternatives die out. Hypothetically speaking, the very existence of open source alternatives would facilitate the legitimate business case that there is an interesting compete scenario possible that needs to be funded. Of course, such a thing is in my mind at this time is completely guesswork, but it seems like a plausible thing to me ;)