gw,
notebooks, notebooks, notebooks
I would say servers first, but overall, I think it is ASPs, ASPs, ASPs. AMD is doing a great job with K8 ASPs, and the key is to win keep winning sockets where the > $100 price for CPU is acceptable.
Then, as a bonus, maybe we can get above 8M unit shipments when 90nm is fully online.
BTW, since I am in a negative mood today, let me say this: 90nm sucks so far. Both from the binning perspective and from wafer starts. These 2 can be related. Why start a whole bunch of wafers if the sweet-spot is 2.1 GHz, while 130nm sweet spot is 2.3 GHz?
90nm wafer start crossover by the end of the quarter means 90nm crossover for parts sold a quarter later, which is end of Q1. The bottom line is that the much disputed one quarter delay of 90nm has happened, and it is not at all clear if D0 stepping can catch up with CG stepping anytime soon, or at all, or if we will have to wait for E0 stepping, which may be well into Q2 2005 (WAG).
Joe |
