X
Business

How Intel calls the tune

In his inaugural column for ZDNN, long-time industry veteran John Dickinson raises the curtain on the real constellation of power in the computer business.
Written by John Dickinson, Contributor
COMMENTARY--Conventional wisdom has it that Microsoft's business practices have driven this industry to its collective drink--and there is a good deal of truth to that.

Microsoft's requirements for building Windows-based PCs alone are enough to make a company quit the business, and that's not including the part where Redmond gets testy when you break its rules.

Somehow, the other partner in this Wintel world we live in, escapes the bad guy imprimatur. But, in fact, friendly old Intel wields a mighty big stick in the computer business these days. What's more, many of its seemingly benevolent practices aren't so benevolent if you happen to be in the PC manufacturing business.

The most flagrant example is Intel's insistence on 90-day upgrade cycles for its leading (Pentium) and trailing (Celeron) lines of microprocessors. Every 90 days, Intel upgrades the speed of its microprocessors for laptop and desktop models and for both its Pentium and Celeron lines. There is always a new and faster processor at the high-end, while low-end Pentiums drop into the Celeron line. And the lowest end stuff that drops off the Celeron line--that winds up running elevators or antilock brakes on some fancy cars.

You might think at first blush this is a good thing because users want ever-faster computers. But the incremental performance is usually non-existent or at best, difficult to detect. Also, users are no longer clamoring for faster speeds; things go fast enough these days, thank you, and current budgets won't support a new round of desktop or laptop upgrades, at least not at the highest end of performance.

Step in line--or else
If manufacturers want their chip allocations from Intel, they are stuck building the newest, fastest boxes; these days, that usually means getting stuck with a substantial inventory of older, slower machines. That's not a great position to be in because people shopping for new computers usually decide they might as well get the newest while they're at it. If you don't believe me, just check out IBM's huge closeout presence on eBay or check what PC manufacturers are saying about inventory levels in their quarterly reports.

And lest I forget, there's Intel's pattern when launching a new chip line, such as its latest Pentium IV. The company's PR flaks get people like us to tout the fact that the chips are coming out at no increase in price. Sounds good, maybe even great, right? But it's terrible because the motherboard infrastructure changes that the new chip requires, including controller chipsets made only by Intel, add upwards of $200 to the cost of the computer, something that no one at Intel is willing to talk about.

If the PC manufacturers want to keep on making computers, they have to use the new chips, and probably absorb a good deal of that cost in their prices (They'll of course wind up carrying leftover inventory of machines based on the older chips.) If they don't, Intel won't allocate chips of any variety to them.

Intel has a standard excuse for all this: If we don't push the performance and product envelopes, then the "competition" (read AMD), will. I'm not willing to say there is no way AMD can hurt Intel, but it's just not the case that failing to do 90-day upgrade cycles or to push new chips out with hidden higher costs, is doing anything significant to hold off the renegades from Texas.

The fact is that Intel is exploiting its exclusive--some might call it "monopoly"--position in the microprocessor market. Yes, Intel is crying the earnings blues just like everybody else because this year's PC volumes are down. But the company is not exactly falling apart as are many of its computer-manufacturing customers.

Still, Intel must be careful. If enough of those companies leave the business because they can't clear the inventories that result from accelerated upgrade cycles, the company could lose its hold on the center of gravity of this industry.

And then Intel's leverage would just disappear.

John Dickinson has worked in the computer industry for more than 30 years in positions ranging from systems analyst and software engineer to editor, writer, critic and industry analyst. His most recent engagement was at eMachines, where he managed the company's Internet and software business units.


Editorial standards