Bob is right in saying that Intel is a trading stock and now may not be the time to own it. That doesn't mean that it won't be a good buy again in the future.
Last week I attended the Microprocessor Forum, the yearly conference that discusses the microprocessor industry, mostly from a technical perspective. For the x86 and IA-64 sections, they had presentations from luminaries including Jerry Sanders and Greg Favor from AMD/Nextgen, Robert Colwell, Fred Pollack and John Crawford from Intel, Glenn Henry from IDT/Centaur, Robert Maher from Cyrix, Joel Birnbaum and Jerry Huck from HP. They had a panel of one person from each company which was asked difficult questions by Michael Slater.
The conference provided the following information -
Intel has 90% of the x86 market.
The AMD K6 can not ramp up its production quickly enough to meet demand (not surprising with .25u geometries.)
The 3 x86 clone chip makers have chosen different directions for their MMX2 instruction set This may fragment the marketplace. At least one of them expects application developers to write special code just for their MMX2 instructions.
They also are taking different, potentially incompatible, directions for system bus bandwidth beyond 66Mhz socket 7.
The IA-64 architecture looks very promising by allowing much greater internal parallelism. Merced will be out mid-99 on a 64 bit version of NT. Intel will keep the IA-32 (x86) architecture alive for the low and mid end while using IA-64 for servers. HP's system will run both NT and unix. John Mashey described the presentations as "one-pixel of the Mona Lisa."
Let me pose a few questions -
What happens when Intel releases their 0.25u geometry processors and their die size becomes similar to the K6? Will that reduce their manufacturing costs? If they start competing on process technology, who is more likely to win - Intel or AMD?
Would Intel lower their prices and generous profit margins in the short term to maintain or increase market share?
Can Intel increase their profit margins long term if they split the market into a separately priced client and server chips?
What happens when NT 5 and Win 98 come out and may need more computrons than current machines provide?
How do the clone chip makers continue to increase performance if they are limited by socket 7 and frontend caches? Intel's slots 1 & 2 solve these limits.
How enthusiastic will the chipset and motherboard manufacturers be at making unique parts for 1/3 of 10% of the market?
Intel and Microsoft are the authors of "PC 98" which is the document that specifies the direction for next year's PCs. Isn't it likely that future versions will continue to be influential and continue to be co-authored by Intel?
Will the "Low cost business PC" market be important this year given that NT 4 doesn't have ACPI, good PnP, or NetPC initiatives that will lower the Total Cost of Ownership (TCO.)? These come with NT 5 in the middle of next year.
Which chips are the major box makers using? Are either HP or Compaq likely to change vendors given their stategic relationship with Intel? Hasn't HP been a force behind the NetPC inititive and won't it likely be a major player in the business use PC market?
My conclusion is that Intel will maintain its "monopoly" for a long time and it will have to compete on price in the short term. Of course, the FTC may have something to say.
-- Alan |
