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.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Petz who wrote (63382)	11/11/2001 2:27:01 AM
From: dale_laroy	Read Replies (1) \| Respond to of 275872

>It sounded like the UMC fab would be new. Why does UMC want to build a new fab, especially a 12" one, unless this is a very long term plan. Does this partnership replace the idea of partnering to build a fab jointly?< Forgive me, I've lost track of what was mentioned recently. However, I do know that UMC was building a 300mm wafer fab to begin shipments from in H1 2003. I also know that they were looking for a partner. Perhaps AMD will become the partner in this 300mm UMC fab?

To: Petz who wrote (63382)	11/11/2001 9:00:09 AM
From: combjelly	Read Replies (2) \| Respond to of 275872

" Why does UMC want to build a new fab, especially a 12" one, unless this is a very long term plan." The whole rationale behind contract fabs is that they can get their economy of scale by running lots of little jobs. So cost per die is a very important point. This is why foundry processes are tuned for yield and not performance, although, to be fair, most of their customers really don't need all that high of a performance level. "Does this partnership replace the idea of partnering to build a fab jointly?" Only Jerry and Hector really know the answer to this question. I suspect this is just a natural out-growth of their "virtual gorilla" strategy. Whether it is true or not, AMD seems to feel that a big reason why OEMs won't use them in their corporate lines is that AMD is unable to supply a complete, top to bottom solution. So now AMD has the Duron (mobile, desktop and very soon, MP), Athlon (mobile, desktop and MP) and is developing the Hammers (mobile, desktop, SMP and 2, 4, 8 and NUMA). This way, AMD has products in every category and can be a supplier if an OEM fears sudden, spot shortages of Intel products if they adopt AMD in their corporate lines. Now this is not to say that Intel has been threatening this, but they have been in court over similar issues in the past, so it is a natural concern with any OEM. Yeah, ok, I know I haven't addressed UMC yet. A contract manufacturer is desirable because they can be used to take advantage of those cases when AMD is capacity limited. AMD wound up leaving a bunch of money on the table during 2000 and even during 2001 because they couldn't ramp fast enough. UMC's 0.13 micron process seems more tuned for low power, but that isn't a huge problem considering AMD's apparent acceptance in the mobile market. Hmm, come to think of it, I am guessing that Barton is going to be farmed out for the most part, and Dresden will be dedicated for the most part to Hammers. Looking at JC's summary of the available speed grades, the mobile processors take a Model jump from Q2 to Q3 of M1900 to M2200. To put it in perspective, that is likely a speed of 1.8GHz (assuming linear scaling), or the same speed as the P4 Celeron that is supposed to be introduced in the same time frame. Hmm, it out to make a heck of a blade server processor... So maybe the UMC process isn't all that slow. Regardless, I doubt if AMD is going to be relying on UMC to the point that they won't be building new fabs. I suspect that they will be building the fab for 65nm and 300mm wafers, with a partner or not.