re: I don't see how TSMC even gets the half-step to FinFET much less to 10nm without changing its business model or eliminating its dividends.
Seeing how they have already started risk production I believe they're quite happy about 16nm volume in 2015. Samsung seem to be too though they say even less.
10nm certainly is rather more problematic, I can't see them doing risk production till 2016 so that would mean volume in 2017 at best for them. You never can really tell what is happening inside companies about things like this though that far ahead.
As to the whole fab business collapsing like you seem to think - I really don't see that happening. The most that will happen is there will be a couple more mergers and the prices will go up a bit but not so much as to cause any major problems. The thing I see being talked about more is how locked in are the SoC vendors to a particular fab, should they be more locked in to get better performance or less so as that means they can move around easier.