In the silicon industry, true dat. Catch is the tooling, systems and process migration to other facilities that will actually produce the end product for a manufacturer. Qualcomm should milk 28nm for as long as they can, since sweet margin once all the capital has been absorbed with return on investment.
This is the part of the discussion before where you and I differ in semantics. There is a product roadmap that will fit in regards to the tech available, but Qualcomm will balance their P&L based on that and what their customers want. The combination of course equals a diminished return curve that any business should manage through to maximize profit.
Simply put, they will milk 28nm for as long as their customers will let them and until 20nm is an efficient process in their own facilities. Now, if you are saying Qualcomm is ready now and are moving forward, inside information always trumps
I place bets that we will not see the 810 chip in many devices (mass) until next spring, but if Qualcomm is ready now, perhaps a shift of production to afford some room for 810 chips later this year? Cool if correct!