Xeon profit margins are too good to be true, read about trouble in server-dom:
zdnet.com
Intel has lowered prices on its Pentium II Xeon processor
before the new chip has even shipped. Intel had originally
set Xeon prices as high as $4,500, drawing criticism from
OEMs. But when the top-of-the-line Xeon -- a 450MHz
processor with 2MB of cache -- ships next month, it will be
priced at $3,690, sources said.
In some cases, Intel will drop prices of 450MHz Xeon chips
to the same level as 400MHz processors, presaging a short
shelf life for the 400MHz chips and machines built around
them.
As a result, several workstation makers are waiting until
early fall for the 450MHz version before they release their
first Xeon-based systems.
"Why should we come out with a 400[MHz system] when
Intel's going to phase [the chip] out in favor of the 450MHz
version three months later?" said an executive with a major
computer company, who requested anonymity. (I wonder why!)
If they sell a half million of these still-overpriced chips, at an average $500 additional profit above the standard P2-400, they will still make 250,000,000 on them. However, if you figure the R&D for the Xeon at 10% of Intel's projected '98 R&D ($2.8B), they have to cover $280,000,000 with it, a loss of $30M on the investment.
For the K6-2, I figure that it cost AMD one full quarter of R&D to develop it, thats $128M in the last quarter. AMD will sell about 8,000,000 K6-2's this year. Since they make $24 more per chip ($27 higher ASP minus $3 additional expenses), the additional 192M gross profit easily covers the 128M R&D in the first year of production.
Petz |
