David Coursey: Intel no longer dominates the PC landscape the way it once did. But that doesn't mean the company's irrelevant...yet. Here's what I think the company's got to do to stay in the game.

I'm still waiting for someone to tell me why I need a 3GHz processor -- or, for that matter, a 2GHz CPU. Sure, Windows XP likes raw horsepower. But my impression is that the OS, if forced to choose, would rather have more memory. And what slows down Microsoft Office, at least for me, isn't the processor -- it's the time required for me to think of what I want to do.

And, sure, if I were a hardcore gamer, for whom a hot-rod box from Alienware, Velocity, or Falcon would be an object of tremendous desire, I might feel differently. But I'm not, and the faster Intel makes processors, the less I seem to care.

Which raises the question: is Intel irrelevant?

During the last half of the 1990s, Intel tried to create new uses for its processors, pushing technologies and products that would eat as many processor cycles as possible. The company was particularly interested in all things video, seeing the editing of home movies as a task that would convince home users to replace their old PCs with something faster. Intel also hoped broadband would drive demand for more CPU cycles. Neither has panned out the way the company might have hoped, at least not in the sense of boosting CPU sales.

The software publishers, what's left of them, haven't helped much, either. Besides games, it's hard to find an application that really requires a faster machine than most of us had three or four years ago. That's particularly true in business, where many 500MHz PCs are still chugging along quite happily -- often running Windows 98.

So Intel seems to have given up -- at least temporarily -- on the search for some fundamental technology that will really change things. Instead, it seems intent on plundering other companies' revenue streams by making its chips do more of the work.

The Intel that presented itself at last week's Intel Developer Forum seems to be one that has given up the search for a fantasy future of 100Mbps broadband everywhere and applications so sophisticated that even a 3GHz CPU wouldn't be enough and has settled into the hard reality: the best way to sell more stuff is to build more of it onto the motherboard.

Page Two

Take, for example, the new Centrino mobile chipset. Every time I hear the word "Centrino" I imagine some silly new atomic particle and start laughing. But if you're in the radio frequency (RF) business, this is deadly serious. One hardware manufacturer has told me that he's going to end up making Centrino-based notebooks whether he likes it or not -- in order to remain in Intel's good graces -- and can only hope the 802.11b wireless that comes with the chipset actually works well.

I'm always sceptical when Intel ventures off into something outside its core competencies, which I've never thought included RF design. And, had Intel been interested in pushing the envelope, it would have adopted the faster, yet compatible, 802.11g standard for Centrino, as Apple recently did with its new AirPort Extreme. By adopting the slower 802.11b standard instead, Intel is following the marketplace at a safe distance.

This being Intel, I have no doubt that its wireless technology has been done right. We'll find out for sure in a month or so after the first Centrino-based portables are released.

The point is that Intel calls its current state of affairs "being in the solutions business," or packaging things together that used to be purchased separately. This is the story of PC hardware, where almost everything you used to purchase separately -- things like modems, network adapters, and video cards -- are now included with the PC itself.

Centrino promises to make third-party wireless add-ons a thing of the past, at least if Intel has anything to say about it. And while it's only to be expected that what once were add-ons will eventually find their way into the basic chipset from which machines are assembled, this strategy only demonstrates how unsuccessful Intel has been in really pushing the envelope.

Rather than forcing progress, Centrino has the effect of homogenising hardware, and removing wireless performance and, perhaps, battery life as areas for competitive innovation. And while this may be inevitable, it leaves fewer and fewer areas where hardware manufacturers can actually compete. And it positions Intel not as an innovator, but as a follower.

Copyright © 1995-2008 CNET Networks, Inc. All rights reserved
ZDNET is a registered service mark of CNET Networks, Inc. ZDNET Logo is a service mark of CNET Networks, Inc.