AMD leaves Intel clutching at CPU pricing straws

IN THE PROCESSOR field, there have been many power shifts over the last few decades. It used to be that RISC chips were the king of the roost, but recently x86 variants are doing more than nipping at their heels. In the x86 world, there was only one name for years and years, Intel. All others copied and tweaked, but few came close. If anyone did, Intel dropped its prices a bit, hitting the competition with the price hammer.

This worked quite well. Too well, and the competitors dropped from many to one, AMD, but this was still a mere shadow of a threat. Then came the Athlon. AMD priced these chips to be faster and cheaper than the competitive Pentium IIIs, and made marketshare gains. Intel came out with faster chips, dipping into the great jar of "reserve speed grades", but it wasn't enough. AMD took the lead for a bit, and Intel tried the price hammer. Didn't work as well as it was hoped, but good enough to make AMD hurt. Intel stumbled with Rambus, allowing AMD to become established as a credible player.

Then AMD shot itself in the foot again and again. Palamino and TBredA were more than enough to allow Intel to regain the lead. This allowed it to fatten margins, put things in reserve, and start swinging the price hammer again, with great effect. Intel was firmly in the lead, and AMD was floundering.

At IDF this fall, between the time that we pre-announced the P4EE and the time Mr. Burns post-us announced it, I had a meeting with AMD. One of the first things out of my mouth was "Guess what Intel just did?", and told them about the P4EE. The AMD engineer just shrugged and nonchalantly said "We expect to have the performance lead on Tuesday". He was grinning, he knew what AMD had.

Current benchmarks show that the FX51 is the gaming performance leader among chips you can buy now. When the EE comes out in the next few weeks or so, it will make up most of the ground, and win a few benchmarks here and there, but in my opinion, it is not enough to be a "leader". Whatever the case, the FX53 will be out in the not too distant future, and that will be the leader, period.

Until the much delayed Prescott hits, there is nothing Intel can pull out of its many hats, and should it make a P4EE/3.4, that will put Presccott in quite a pickle, won't it?

That probably won't happen, mainly because the price of the current (future?) EE is in the "don't buy me" range, $925. Intel product planners must be buying antacids in crates.

So that said, for the next three months or so, Intel has no response, and the launch Prescotts don't look able to bring the performance crown, much less do it convincingly. AMD knows this, and acting accordingly with the first round is the price cuts it didn't officially make last Friday.

Intel dropped the prices of the P4/3.2 from $637 to $417, a gesture usually accompanied by a new processor introduction. When Intel does this without a new slice of silicon, it is not a gesture of goodwill, or an indication that the trucks that vacuum the wads of cash out of the Intel vaults are late this week, it the price hammer. If Intel had the lead in mindshare or raw speed, which it pretty much always has had, AMD has to respond. It always did like good little followers.

Friday, Intel cut, and on everything that mattered, the Athlon64s and top Opterons, AMD didn't change prices a bit. Shortages abound on the FX chips, but several resellers tell me that they are not having problems with 64/3200+ availability. This points toward AMD deftly avoiding the price hammer strike, and doing what it wants. That is more telling than any published roadmap now isn't it. µ

See Also
Is Intel Yamhill to be AMD incompatible?
Intel's Prescott may be delayed until Q2 2004
AMD holds firm on pricing, as Intel makes desktop cuts
Intel Prescott to be delayed until January, February - Extreme Edition to cost $925
Up to date Intel roadmaps