But surely you do not disagree with the original claim that TSMC wins over GloFo are bad for AMD and good for Intel?
Surely TSMC wins using Intel IP are good for Intel.
Not convinced that there isn't enough demand, long-term, for cutting edge GF foundry development. Like I said, the dynamics of the market have changed; Even if AMD had not spun off GF, the dynamics would be utterly different right now for the same reasons I described... Being freed of the production outsourcing constraints of the old x-licensing agreement, i.e being able to fab without limit at Chartered, releases a lot of pressure on GF/Dresden. I think AMD can immediately enjoy the luxury of changing to a more boutique product mix at GF/Dresden, while (at least) making up the difference at GF/Chartered.
Pseudo-graphically (hope this comes out right (nope, no monospaced tag)),
Fab: GF/Dresden|GF/Chartered
'05: ====-----
'06: ====------|-
'07: ===-------|--
'08: ====------|--
'09: ====------|--
'10: =======---|----------
'11: ++++======|====------------
Where "=" is a boutique part, while "-" is a mainstream part, and "+" is a future ++boutique++ part, designed specifically to translate high-performance capacity into high-margin parts.
IMHO this is (potentially) a MAJOR shift in the dynamics of the market; For several reasons, the old thinking that Intel can throw its weight around and squash AMD like a bug may be obsolete.
fpg
AI
Market Data 
Markets 
