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: Gopher Broke who wrote (63338)	11/10/2001 1:21:00 PM
From: pgerassi	Read Replies (1) \| Respond to of 275872

Dear Gopher: The fact that P4 in certain speed grades is sold out leads to three possible conclusions: 1) Sales are far higher than expected. Only the lag time is causing a shortfall. In 3 months (the typical lag time) things should be real rosy. This is the one you are assuming? A few strikes against this is the current shortfalls of Xeons both P4 and large cache P3 variants. These usually are ordered a few months in advance. Thus, the lag should be quite small, if not irrelevant. This points to production (yield) or a marginal design (ala K6 speed grade flaw). Not having enough production lines can not be the problem as Intel has many 0.18u fabs. 2) Production switchover problems. Not enough old lines left that can produce enough P4s. Less rosy as the hope would be that they could be filled from 130nm NW lines later. This is deprecated as Intel supposedly has many 180nm fabs, so this is probably not the problem. Another could be that the 423 pins P4s are not being ordered and not enough MBs are available. Not so, according to the Taiwanese MB makers. 3) Yield problems. Not enough of the CPUs are binning high enough to use. They are runnable, but not saleable. This explains the the many other facts. High speed bins are not available and are sold out. This would not be fixed by over producing wafers as the CPU costs go way up for no additional revenue (ala K6 speed flaw). The bin yield curve is lower than the pricing curve would demand (ie, marketing expected very little yield of 1.3 to 1.7GHz and packaged for 478 pin P4s when the grades below that offered at 478 pin packages would need to be in 423 pin packages). This is the down side of two different package types coexisting and the fact that Intel would be embarrassed to show that the slower bins also use the newer package (proving the speed curve is lower than expected). To keep from showing this fact, Intel would rather throw away the CPUs. This fits the outside facts more closely and is therefore, the right read on the situation. It also does not look like something that will be solved quickly. Each attempt needs three months to prove itself. It only takes one or two misses to be a really bad problem. Pete