Intel's stock has soared so high for so long, it may seem like no big deal that it has come down a bit. As a customer, you certainly don't care whether the company's disappointing quarter was because of slowing PC demand, channel inventory adjustments or the impact of cheap Asian currencies. But occasionally there are deeper explanations that are worth understanding.
This is one of those times.
For the past 18 months, I've been saying that until at least the turn of the century, network computers will be much more of a metaphor than a market. This has never been more true than it is right now. While network computers from IBM, Sun Microsystems and Oracle continue to have just about zero market impact, the power of network-centric computing is changing the very nature of the PC business.
You have a good sense by now of what network computers are all about. Desktop computers are more like network access vehicles than separate computing platforms, and IT investments have shifted toward the enterprise network infrastructure. The logical consequence is that the overall pace of PC upgrades should slow, and when you do buy additional PCs, you will likely trade lower performance for lower prices. Those lower prices will further the long-term goal of device ubiquity, despite today's apparent unit demand blip.
From Intel's perspective, that is a mixed blessing. Since the company is the world's only high-volume, Microsoft-compatible chip manufacturer, the more PCs the better. On the other hand, Intel has always benefited enormously from customers' willingness to adopt its latest and fastest microprocessors. That mutual commitment is fading away.
Remember how employees used to beg, battle and barter for advanced 286-, 386- and 486-based systems? That meant customers and Intel were on the same upgrade cycle - a concept consistent with a PC-centric view of computing. But in a network-centric world, the gap between what Intel can provide and what customers need will steadily widen unless a whole new range of desktop applications emerges. That's why Intel is emphasising audio, video and 3-D technology.
But that strategy isn't working yet. Intel made a big deal about its MMX multimedia technology, but what has been the real customer impact? Now it is rolling out a low-end brand called Celeron. This will be tricky. Andy Grove and company have spent the past few years telling us we are nowhere if we don't have a Pentium. Now that the low end is by far the fastest-growing PC segment, that message apparently will change fundamentally.
Competition is critical to this whole process. As a customer, you may not think much nor often about AMD and Cyrix. They remain minor players that have often found it hard to keep their promises. But if it weren't for their perseverance, imagine what Intel's prices would be like today. Barrack for them. Like Amdahl, they are probably destined to save customers far more money than they ever earn for themselves.
Don't get me wrong: Intel's future is still incredible. Once its 64-bit Merced chip is available, it will take over the server market for Windows NT and Unix-type systems - an unprecedented achievement. And someday, we will almost surely want our computers to speak, hear and see. But in the meantime, PCs are looking more and more like network appendages. And as long as that is the case, falling prices will be the spirit of the times.
(Moschella is an author, independent consultant and weekly columnist for Computerworld. His e-mail address is firstname.lastname@example.org)