I've thought they're stuffed for years
My opinion (which hasn't really been proved either way yet) has been that Microsoft's biggest problem is the fact that they need to keep the revenues going for the shareholders - that means keeping selling all those CALs that don't make much sense any more, the difficulty of developing new techniques without impacting existing cash cows - hence the netbook debacle with XP having to be revived and the question of how much should a web-enabled office be able to do - and the forced "innovation" in Windows so that there's something to sell every couple of years for the usual range of price tags.
This results in a licensing headache for companies that the "Microsoft ecosystem" tries to fix, but the problem really needs to be tackled at source.
Simplification is what's needed, but that is undoubtedly going to mean ripping up their existing licensing and development playbook and sacrifcing short-term cashflow which the shareholders might not like so much.
The question is how long the slide has to continue - will they wait until they've lost enough market share across all markets that the switch is easier because fewer customers are affected, or will they take the hit and transform into something relevant before that slide reaches a tipping point.
I think the answer to that question will determine the relevance and success of Microsoft in the future.