Klunk

Got this email as my wake-up call this morning, entitled, “2200 $45 one day price drop.”

“I’d like to know what’s behind this. I read quite a few sites everyday and can’t recall anyone of you guys calling this one. . . . I’d be super pissed if I bought one sunday, only to find a 25% price drop the next day.

I go over to Pricewatch, and sure enough, the price of the 2200+ has dropped quite a bit, down to $155 with shipping at Googlegear, for instance.

So why wasn’t this called? Well, we did mention a few days ago that the prices on the upper-rated Athlon prices would likely fall quite a bit as the result of the Labor Day price cuts by Intel, but it’s hardly Labor Day.

Pricing: Official vs. Real

It used to be that official pricing meant something, and that the price you paid for a processor was the official price, plus a little bit.

Those days are gone. Nowadays, the price you pay for an AMD CPU rarely bears any resemblance to its “official” price; the low Pricewatch price is usually 33-50% off the “official” price.

Intel prices have also wandered from the price reservation, though not nearly as much. The Pricewatch price on an Intel processor will usually be a little below the official price, but pretty close.

Except when it isn’t. For instance, the 2.4 and 2.53 are selling at about a 15% discount from the official price.

The timing and mechanisms of price cuts also differs quite a bit between the two companies.

Intel is usually pretty methodical and predictable. They announce products/price cuts at certain points in the year. If it’s around Martin Luther King Day (Jan. 15), U.S. income tax day (April 15), Memorial Day (end of May), Labor Day (end of August/early September), or (less likely) Halloween (end of October); it’s a pretty safe bet Intel is going to do something with products and/or pricing.

Even their “surprises” are methodical. For product introduction, Intel’s idea of a surprise is to push an announcement up by a quarter; I don’t recall a single exception to this the last few years.

They usually do the same with prices, though they’ll break that rule more often. For instance, the pricing on the 2.4B will be $193 after Labor Day rather than the $241 expected from the roadmaps.

AMD is a lot less methodical and predictable. They do like to announce major products at the very end of a quarter, but not always.

Pricing is very unpredictable. While AMD will almost always “officially” announce price cuts right after Intel; what they’ve gotten into the habit of doing lately is to quietly cut the price without announcement before Intel acts. From what I’ve heard, even resellers don’t get much warning on this.

So what is becoming the norm for AMD is price cuts without announcements, and announcements without price cuts.

This is the most likely explanation for the cuts on the 2200+, though this is a bit earlier than we’ve seen them before.

There are two other possible reasons for this particular cut:

TBred Isn’t Selling

I asked resellers a little while back about TBred sales, and those who wrote me said that the 2200+ was not selling terribly well. They said that OEMs were shying away from the processor due to difficulties in cheaply cooling it, and they also said that people were buying the much cheaper Palominos.

A good rule-of-thumb on AMD pricing is that the pricing on an AMD processor should be at least a third less than the Intel equivalent for it to sell. That’s pretty much been the pricing pattern the last couple years. By that standard, the 2200+ was overpriced. Matters were made even worse by clearance prices on Palominos (which have been selling pretty briskly at prices about half what equivalent Intel chips cost).

The real problem AMD faces after Labor Day is the price of the 2.4GHz PIV. At $190, it seems to limit the price of any 2400+ to about $130, which is hardly going to help AMD’s financial results.

This is why AMD needs a 2600+ and especially a 2800+ as soon as possible; they’ll get decent money for a 2600+ and pretty good money from a 2800+. AMD needs that kind of money now when they’re probably pocket less than $60 for most of the Athlon sales and $40 for Duron sales.

A 2400+ Real Soon?

Finally, there is the possibility the 2200+ price cut is a precursor for perhaps a pre-Labor Day introduction of the 2400+. As stated earlier, the problem with that is the imminent $190 2.4GHz PIV price. It’s better than nothing, but only points the need for faster processors to help out the average selling price, which was in dire straits last quarter and will probably be worse this quarter.

AMD’s Prospects?

Geeks aren’t too strong on finances (just look at the dot.coms), but it will be the financial markets that in the end that will determine AMD’s fate.

Two-three years ago, I don’t even write articles like this. People would be begging AMD to take their money to tide them over.

That was then, this is now. Right now, the financial markets are pretty hostile towards tech firms, period, much less those that are down on their luck.

God help AMD if it’s found they’ve been cooking the books, and if I were the SEC, I’d be rather interested in AMD’s inventory accounting right now.

Nonetheless, even without financial finagling, AMD cannot indefinitely lose $200 million a quarter. They cannot indefinitely only make $50-60 a processor. The financing and operation of Dresden has added quite a bit to the bills they have to pay each quarter. They need at least $600 million in revenues from the CPU industry to break even. Last quarter, they got less than $400 million. Next quarter looks even worse. Christmas quarter may mean more sales, but price per processor will still remain low. They have about $800 million in the bank to pay for losses.

Unless something wonderful happens, by New Year’s Day, odds are AMD will have something more like $400 million left in the bank. Intel’s aggressive pricing combined with AMD’s inventory problems make it difficult to see how they get themselves out of the hole they’re in for the remainder of the year.

Then the vultures start hovering above, and banks stop returning your phone calls.

Then it becomes a race between Hammer and the balance sheet. That will be the biggest computer hardware story in 2003: Will Hammer Save AMD?

Yes, then Hammer is supposed to make everything all better, and it well may, but even the best processors take time to be accepted, and a lot of time is not what AMD is going to have. Unless they can get a hell of a lot more out of the Athlon design than we’ve seen so far, it’s not going to be much of a financial prop paying the bills while Hammer tries to be a hit.

If AMD can sell a million Clawhammers at $250 each that first quarter, or one way or the other get a couple hundred million revenues from them in the first, or at latest the second quarter, then AMD can beat the vultures off.

If they don’t, then either they find a white knight (that’s somebody who either pumps money into the company or has a friendly takeover) to tide them over, they get taken over, or (least likely) they just go out of business.

What do I think the odds are on the first two happening? I’d say 30% at this point.

Somebody will certainly want to buy AMD, or at least its assets. But who?

nVidia might have been a candidate six months ago, but their stock price has shrivelled. They just don’t have the money.

IBM is trying to get out of the hardware business, not into it.

HP/Compaq is a possiblity (though a more likely white knight than a buyer). The Taiwanese foundry folks might be interested.

But which company is the one that would most benefit from AMD ceasing to exist as it is now? The only one that would still benefit even if they bought every AMD asset just to blow it up?

Yes, Intel would have some serious antitrust problems, but tell Dubya that a lot of Texas jobs would get saved, and either Mr. Schroder or Steiber (next year’s possible German chancellors) that even more Dresden jobs would be saved, and this might not be impossible to do.

Not saying it will happen, but if there will ever be an opportunity for Intel to do the one thing that will dramatically improve its prospects and profits, the upcoming year will be the time.

To The AMDroids

You don’t like this possibility? I can’t say I like it very much either, but things don’t look all too good, and saying nothing doesn’t make the problem go away.

Want to make the problem go away? Start buying the top-end AMD processor every month or two. At the least, build three Hammer systems the first half of the year.

Support your team and put your money where your mouth is.

Ed