Intel, AMD: Countdown to a silicon Waterloo

"We will do what we need to do to transition from the Pentium III to the Pentium 4" Paul Otellini, general manager of the Intel Architecture Group, said during the company's second-quarter conference call. "I would categorize this as an aggressive transition as Intel has driven since the 486 to Pentium transition."

Otellini, the presumptive heir apparent to CEO Craig Barrett, has mastered the Intel dialect, where executives dispense opaque generalities and yet simultaneously convey specific factual information about what happens next.

In early 1999, for example, a reporter asked if free PCs would appear. Otellini said it was not likely, but added there would be "a lot of interesting experimentation" in PC pricing. Months later, CompuServe and others began offering $400 rebates.

At last year's Comdex show, he dropped a line about how consumers will increasingly incorporate video in PC communication. "He's high," thought the collective pack of hungover reporters and analysts. "There's not enough bandwidth." Eight months later, Microsoft and Yahoo are talking up video instant messaging.

The most recent unmistakable nugget, of course, is that the company will continue its scorched earth pricing policy. In true Intel tradition, of course, Otellini never banged his shoe, Khruschev-like, and bellowed "We will bury you!" He's more likely to wear lederhosen at the next stockholder's meeting than raise his voice. He never even mentioned AMD by name.

He didn't need to; the war between these two rivals often follows a familiar pattern: AMD introduces a chip, gains market share, and then gets steamrolled in a price war that Intel claims isn't going on.

In the second quarter of 1998, AMD began to emerge from the 1997 war. By the fourth quarter, AMD controlled 16 percent of the entire market for Windows computers, double the year before.

Then came the veiled threat.

In early 1999, Intel unveiled new Celerons and talked about an upcoming chipset with integrated graphics that would cut manufacturing costs.

"No matter how much we talk about performance, the only thing that matters is megahertz," Otellini murmured. "We will go with the natural flow this time."

A round of then-unprecedented Intel price cuts and chips introductions followed. AMD's revenue dropped from $788.8 million in the fourth quarter to $631.6 million while growing profits tumbled to a $22 million loss.

In the next two quarters, AMD lost more than $100 million.

In the current situation, AMD made tremendous strides in 2000 with the advent of Athlon and Duron. For the first time in five years, annual profitability returned as market share grew.

Aftermath of price war
Then the price cuts started in April. The megahertz gap between Intel and AMD processors also widened. Revenue, profits and processor prices declined in the second quarter for AMD. The company says it may lose money in the third quarter.

"The decline in ASPs (average selling prices) was influenced by two factors," the company said. "Aggressive price cuts that became effective during the quarter, and a lower proportion of 233 megahertz devices than planned."

Oops. That was the 1997 note. This year, AMD said, "An unprecedented number of price moves in the quarter by Intel in attempt to stall our market share progress were successfully countered but impacted ASPs severely."

Intel's advantage in a price war lay partly in its greater factory capacity. With 13-plus fabrication facilities, or fabs, Intel can simply produce far more chips than AMD can with two fabs within a given amount of time. That not only lets Intel produce cheaper products, other circumstances being equal--it also offers latitude to boost speeds faster.

For PC makers facing declining returns, cheap rules. Pile on the volume discounts, technology licenses and advertising dollars Intel can dole out, and AMD quickly becomes mathematically unattractive.

"In an environment like we have today, you don't see too many people sticking their heads up and trying to do something new and aggressive," AMD CEO Jerry Sanders said during the company's conference call.

Last year at this time, Sanders predicted major PC makers would adopt Athlon for the U.S. corporate market by late 2000. It hasn't happened yet.

Finally, Intel can withstand a price war better because of size.

The circumstances this time around, however, are different. For one thing, the consumer segment has embraced Athlon. AMD has also shed its reputation for manufacturing mishaps. The Athlon also costs less to manufacture, although improvements to the Pentium 4 should close the gap a bit.

The company also has numerous potential markets to penetrate. Although Gateway uses the Athlon, it has yet to seriously adopt Duron. Notebooks are ripe for colonization.

So what happens next? The next round of price cuts take effect on August 26, just in time for back-to-school buying. Microsoft and the PC powers will also release new systems in October to coincide with Windows XP. From now until then, executives from both companies will chug meeting coffee, flip PowerPoint slides, and try to cut as many deals as they humanly can. Never will so few work so hard to insert the word "value" into so many conversations.

If you begin to see an increase in AMD computers in the Sunday newspaper inserts, Intel will have failed. But if you see it go the other way, history will be repeating itself.