ALTR said that even in case TSMC gets FinFet working Intel/ALTR would still have a die size (shrink) advantage compared to TSMC/XLNX.
Intel will have great cost advantage relative to TSMC/Samsung/Glofo FinFets because Intel seems to be able to shrink (indirectly indicated by ALTR) .
Do the analysts figure this one out?
I am not convinced that TSMC will easily get FinFet working. Intel played with FinFet but it has structural issues as you shrink. TSMC is right now engaged in a very risky catch up game with Intel by trying to push a die shrink before they have completely tested it. Even if they manage to get it working I doubt that they will get the yields they want. TSMC thinks they will have 14nm process by 2015 but Intel will be mass producing by 2014. So even by TSMC overly optimistic projections they are still a year behind.
Of course Walliweaver. I remember the pain Intel went through when they came up with tri-gate. Tri gate has its roots in FinFet but the implementation of it is so different that Intel refuses to call their Tri gate process FinFet. Going 3D raises the complication to a whole new level. I remember all the hecklers when Intel was struggling with tri gate. Now I am going to watch TSMC squirm.
Intel put ten years into developing FinFET. The idea that the ARM foundries are just going to crank it out are really pretty silly - especially given how much trouble they have had with just planar on recent nodes...