Rivals move in as Intel falters

The last few months have not been kind to Intel. While all processor manufacturers have faced certain difficulties, such as supplying a hungry market in sufficient volume, Intel seems to be in a hole all of its own making.

Following numerous embarrassing slip-ups, it has started to develop an unhealthy reputation as a company unable to deliver on its promises. A credibility gap has started to open up between Intel's roadmap and what it has actually been able to bring to market, and it is a gap that rivals have been quick to fill.

AMD in particular has exploited Intel's troubles to change its image from financially flaky underdog and second source supplier of processors to the consumer end of the PC market, to confident contender in the corporate PC market with designs on the high end. The launch last week of its new Athlon Thunderbird chip is almost certainly going to cement this position and drive the company onwards into new markets. And the forthcoming Duron processor may well be the next phase in what AMD marketing director Richard Baker calls "our big push into the business market".

Corporate quandary
Transmeta's Crusoe chip has also come out of the blue to pose a threat in the same part of the market as Athlon. Backed by Linux creator Linus Torvalds, Crusoe is mirroring the open source operating system's progress as an increasingly serious alternative to Wintel in the enterprise sector.

The situation presents corporate end users with a quandary. Their PC and server choice is no longer simply between Intel and rival architectures like Sun, Alpha and RS/6000, but includes other contenders cut from quite a different cloth.

Analyst Clive Longbottom said: "For the first time, corporate buyers are able to choose between AMD and Intel on an equal footing, with AMD offering not only a fast machine but a price differential. If you are buying 1000 PCs, we're talking a big price differential."

Spiralling development
The significance of AMD's rise is about a lot more than cheap desktop boxes. "Against what I predicted when Athlon was announced, AMD is starting to look like a high-end contender," said Longbottom. "It would like to get into the server and workstation markets, and while it still has a few issues to sort out before it's there, there's no reason to suppose it can't sustain its current run of success and achieve what it wants."

Longbottom, like many other observers, believes the serious competition that Intel now faces is great news for users, but he has some reservations. "Users benefit price wise from this war, but lose from the speed at which the market is now changing," he said. "Processors have more than doubled in clock speed in the last year, and development, fuelled by competition, is moving too fast for most companies to keep up with. And do we really need all that speed?"

Analysts have criticised AMD and Intel alike for putting marketing ahead of considering what customers really want. The news last week that the University of Kentucky has built "the world's cheapest supercomputer" by clustering 64 Athlon chips into one box will not help. A magnificent achievement it may be, academically speaking, but it is hardly the kind of gimmick that is going to impress hard-headed system specifiers.

Waiting game
This may be turning into an annus horriblis for Intel, but there will be a substantial part of its user base that is prepared to wait until it recovers, and won't be tempted by any upstart alternatives. But how long will all this take? And will it be worth the wait?

Some of the problems dogging the company are beyond its control, but are not necessarily terminal. Rambus dynamic RAM, for instance, a key constituent of its Timna integrated chipsets, is in desperately short supply and remains extremely expensive.

This problem could ease before the end of the year, but other issues look to be casting a longer shadow. For example, in its haste to reassert itself as a leader in technology, Intel has promised its so-called Willamette processor for later this year. It will be the company's first new brand since the Pentium Pro in 1995, and will have a clock speed of at least 2Ghz.

Losing the plot
Leaving aside the issue of whether this clock speed is actually what users want, there is the very real possibility that Willamette is not actually what it appears to be. Intel is apparently readying a successor to take over from Willamette within a few months of its release.

It will be obsolete before it can be developed properly and buyers can place their orders. Moreover, the Willamette is apparently set to be more expensive to manufacture than any Pentium, owing to its greater size. So who on earth is going to buy it?

Even those not especially bothered about Intel's PC processor technology may be concerned that such an influential player in the industry has, at some important level, lost the plot. It is still one of the largest technology companies on the planet, but whether its wares remain the key strategic infrastructure of choice for ever is now open to question.