Here are some of my thoughts (btw, I admit that I am an Intel bull just "dropping in"):
1. Jerry is not stupid
2. Jerry is from the "old school", and does not believe in being forthright or candid, although he probably does not himself believe that he is being dishonest, misleading or deceitful. Good intentions or not, however, as an analyst, OEM or an investor, Jerry cannot be trusted to give you the full, accurate and complete picture unless it is favorable. It's just not his nature to 'fess up about the dark side.
3. Given item 2, if Jerry says there is a yield problem, you had better believe that there is indeed a yield problem.
4. The K-6 is about the same as a Pentium MMX, perhaps even a tad faster, but the difference is insignificant, and the K-6 is no match for a Pentium II.
5. Regarding item 4, however, dominating the CPU market is NOT primarily about technology, which is just one of many factors leading to success. This is a key point that an awful lot of people simply either cannot understand or cannot accept.
6. In the K-6, AMD has a viable point-solution product, but not a comprehensive product line or, for that matter, a long-range strategy.
7. Large losses are unacceptable; but so are small losses, breaking even, and small profits. Within two years, AMD is going to have to start paying interest on the debt to retool the Austin Fab PLUS build the German fab, whose output, by the way, is not certain to have buyers. BIG TROUBLE in two years if Jerry can't turn a BIG profit.
8. In the mean time, Jerry has to now retool again for .25 micron at Austin, and keep up with Intel, which is a moving target.
9. Intel does NOT have yield problems, 300 MHz P IIs are readily available (if expensive), 440 LX is shipping, Intel showed a 350 MHz processor with a 100MHz bus this week, it will be shipping by 2nd quarter next year, possibly months sooner. Intel also has a full product line including notebooks and servers, plus a complete strategic roadmap covering the entire spectrum of CPUs, as well as support chips.
10. Jerry cannot compete with any of the P II chips, he does not have sockets I or II (patented), he does not have the technology of P II (either design or manufacturing), and he does not have a 100 MHz bus.
11. Just about the time that Germany comes on-line, Intel will probably be above 400 MHz w/ 100MHz bus on P II, and will be introducing Merced at 600 to 1,000 MHz. AMD can't compete with Merced, either.
Jerry started a war that he couln't win; he ALMOST won the opening battle (it sure looked good in late March and early April), but the war is probably lost. To make matters worse, he "bet the company" with the cost of building the German fab. The interest payments begin coming due in 1999, I think, about the time that the situation will probably be at its absolute worst. And about the time that Jerry is set to retire. |
