implications the Prescott failure
As long as AMD has the advantage in both power and performance, I think they can take 2% market share per quarter. That is about what they did per quarter when Athlons were outclocking P3s and outperforming early P4s as market share went from about 15% to 23% in 4 quarters.
For Q1 the extra market share and at least constant ASPs are already in the bag. AMD's CPU revenues should grow at least 5% as Intel's fall 8%.
For Q2 Intel may start to solve the Prescott problem but AMD will at some point have the advantage of superior security with SP2 release. So another 2% gain is likely. ASP will probably drop a little (seasonality thing), and total units sold in Q2 is usually 10% lower than Q1, so CPU revs will drop slightly.
For Q3, Intel will probably have solved the Prescott problem, but AMD will maintain its lead as 90 nm sales become significant towards the end of the quarter, so I think the 2% is again assured. There is also significant potential upside for AMD in Q3 if Intel has failed to deliver iAMD64 before the end of Q2 as promised. I predict if there is no Nocona by 7/1/04, DELL will fall to AMD.
For Q4, unless AMD has stumbled in 90 nm, Intel will again be way behind in performance and this will last until Tejas appears.
The above has concentrated on desktop and notebook sales, but I don't think AMD can gain more than 2% server market sales per quarter unless Nocona fails to appear before 7/1. The server market is notoriously conservative, and I don't think more than 10% of HP's or IBM's servers will be Opterons in 2004. If Nocona fails to appear, however, AMD can gain up to 3% per quarter market share in servers. 0% if Nocona is really, really good and arrives on time, 3% if it fails to arrive or its performance is poor.
Petz
