Original URL: https://www.theregister.com/1999/06/18/life_not_so_fab/

Life not so fab at Intel

The process that shagged me

By Mike Magee

Posted in On-Prem, 18th June 1999 10:59 GMT

Analysis This writer, on his Vespa zooming to the Twisted Wheel in Manchester in the 1960s, well remembers the time when fab meant something really ace. But life is far from being fab at Intel, it appears. After a relentless and ruthless round of price cutting throughout the first six months of this year, it now appears that the strain is beginning to affect the gravy train. Add to that the lukewarm reception given to the (late) Merced chip, problems with the Coppermine process, the lateness of the Camino 820 chipset and the whole cafuffle over Direct Rambus, Intel is not having the greatest of years. It's perfectly clear why Intel took the pricing decisions it made with the Celeron for the first six months of this year. It was losing market share to rival AMD because the original Celeron was a disaster and it was late to enter the sub-$1,000 arena. But in choosing to slash prices on the Celeron, Intel found itself between the horns of a dilemma. The Celeron is a cut down Pentium II. The Pentium II cost far more than the Celeron. This is not to say that Intel loses money every time it sells a Celeron. According to senior Intel VP Paul Otellini, when he visited London a few weeks ago, the company makes money on all of its microprocessors. The question is how much money Intel makes on these processors, given that its whole existence depends on high (50+) gross margins. The .18 micron Coppermine process is a different kettle of fish. At the Intel Developer Forum earlier this year, Pat Gelsinger, another senior VP at Intel, made it clear that the company would be moving to a .18 micron process as quickly as possible. By February of next year, Gelsinger estimated that three or four of its fabs would have made the move. It is a tricky matter moving from .25 micron to .18 and you can't do it all in one go. Intel certainly managed to keep to its promise and introduced its first .18 micron mobile processor last week. And Gelsinger at IDF noted that such technology switches always meant mobile processors would arrive before desktop processors. The Camino and the Direct Rambus affairs are a different matter. Last week Craig Barrett, Intel's CEO, admitted for the first time that the firm had a back up plan which could include the PC-133 standard. That's a humiliating climbdown, but even huge multinational corporations have to eat humble pie from time to time, whether they be Intel or Coca Cola. We have faithlessly pursued the Merced story for some years now, and we were the first to report that the processor was delayed some two years back. Intel has invested a huge amount of money in its IA-64 project and it has to make it work. Whether it will or not, is a different matter. Otellini told The Register a few weeks back that its IA-32 Willamette process is still firmly on track. Earlier this year we managed to obtain some inside information on the architecture and it looks funky. Being the liquid corporation it is, Intel has very deep pockets and has such a dominant place in the market that it would be astonishing if it did not continue to do well in the future. But it needs to keep the Andy Grove ethos in mind if it wants the gravy train to keep delivering golden doubloons to its shareholders. It's already seen off Cyrix and IBM Microelectronics, IDT and Rise pose little or no threat, and AMD has its own set of problems. Maybe Intel should be worried about the PlayStation II, which is far more likely to pose a threat than the x.86 cloners. Now that really is a fab machine, man. ®