I was under the impression the exact opposite was true. All future products are coming on GF 14nm LPP and the WSA was amended to lessen the financial blow of all the new console contracts being on TSMC's 16nm. I could be wrong but I thought the point was to increase overall fab capacity while not being penalized as much as before for needing to use another process.
The silence on Vega details is really starting to flirt with being dangerous. Some teasers would be good right about now, nothing performance related but something to tell the public waiting that Vega is coming very soon with full details to follow. Hell I'd take a 10 second snippet with a "catch us at E3!" right now.
I don't think that GF LPP is good enough for Vega. Even with the recent Polaris refresh it's pretty obvious that the process won't cut it if AMD is trying to make a competitive high-end GPU. Even with design improvements to allow it to clock higher, it would have to max out the power draw to hit the suspected clocks. AMD needs any edge they can get and I suspect that Samsung's latest generation of their process would give them some extra wiggle room. Even if GF wants to grow their process, now that AMD has a worthwhile CPU, they can actually move product and they're going to need additional capacity for their server chips and APUs that will be coming out in the near future as well.