Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Process Boy who wrote (53897)	4/2/1999 2:22:00 AM
From: Mani1	Read Replies (2) \| Respond to of 1571405

Process Boy Re <<The 30% unit volume you cited, is that the U.S. retail market, or the whole CPU market?>> That was for the market share for all x86 CPU market. AMD has already a much bigger share than that (50%) in the retail market. I remember that in a conference call from few quarters ago, they said their goal is 25% of the over all market by 2001 and they have potential capacity for "more". Keep in mind that fab 25 still has room to expand and Dresden is just starting to ramp, conversion to .18 increase the capacity even more. 30% market share sounds right with fab 25 and 30, if capacity utilized properly and yields are ok. I see no need for AMD to built a new fab and as far as I know they have no plan for one. Mani

To: Process Boy who wrote (53897)	4/2/1999 2:41:00 AM
From: Yougang Xiao	Read Replies (1) \| Respond to of 1571405

Process Boy: <<AMD will have two fabs with competing technologies (Cu vs. Al) within a year. 30 will have an extremely immature technology (Cu) that I believe could take longer to ramp than is publicly projected. I could be wrong, but I'll believe it when I see it. Assuming they do OK with the Cu ramp; they will at some point have to convert 25. During conversion the capacity will be reduced. And at ~185mm2 die size for K7, they'll have to put out a LOT of good yielding wafers to generate 30%(?) unit volume for the whole CPU market. They could use at least one more fab. Two fabs for the market share they apparently are going for leaves ZERO margin for error in manufacturing execution.>> You have very good points here and I believe every word said here. The obvious implication is to follow your suggestion to get that additional one or two fabs. In essence, to turn AMD into a real and physical gorilla. That would be nice. The problem is that AMD cannot afford them. Now here is AMD's solution: to become a "virtual gorilla" Please see Jerry's speech on this subject at amd.com "virtual" thing is basically to share, to be allianced with others for the resources that AMD does not have or can not afford. In the case of MOT alliance, I believe that they have the foundry support clauses (AMD 1998 10K, out in the last few days) built in. Whether MOT alliance and the foundry support can solve the potential problems as you point out remain to be seen, at least, there is a fall back position. << The 30% unit volume you cited, is that the U.S. retail market, or the whole CPU market? I'm not familiar with this particular projection. Could you refer me to where you got it? Was it part of an AMD conference call or something? Does this include servers as well? As I just postulated, AMD will have to execute flawlessly to achieve the umbers you are suggesting with just two fabs.>> The 30% is AMD's goal in late 2000 or early 2001 to reach worldwide CPU market share. This goal has been publicly stated in one of the CCs. The 30% should include server chips as well.