In a sentence, Intel had a lousy quarter; AMD had a terrible one.
There’s not really a lot to say about Intel. Normally, you’d expect a 5% drop in CPU sales; Intel had more like 8%, with slightly
lower ASPs (which are roughly around the $150-$170 level). Chipsets got rapped by lower prices.
The silver lining partially offsetting a bit-worse-than-ho-hum quarter is that the non-processor-related
areas bled less red ink, and Intel is now starting to move to get rid of the real financial dogs.
You usually find out practically nothing about what Intel is doing at these earnings conference calls, they use other channels to get that kind of information out. They’re pretty much financial only.
AMD was quite another story.
They lost about $180 million. The number of processors they shipped dropped 25%, from eight million to six million. The ASP they got dropped from about $75 to about $63. Market share dropped from a bit over 19% to a bit under 17%.
In short, a bit worse, but close to what we predicted a month ago.
What was news, and disturbing news at that, was the announcement of an inventory overhang. Apparently AMD sold way, way more processors to OEMs and resellers in the first quarter than were used. So much so that they actually had to take some back.
Even worse, there’s still an overhang of Palomino XPs out there, one AMD expects to take the third quarter to finally get rid of. More importantly, AMD expects relatively little growth in unit sales in the third quarter (about 5%), so expect market share at best to continue to hover around 17%.
AMD tried to spin this into a virtue, and talked so much about reducing inventory and providing “fresh product,” I thought they were now in the fish business. No one but overclockers are interested in “fresh product.” The only fresh thing resellers and OEMs are want are “fresh prices.”
The low expectations for the third quarter indicate that this is not just a matter of “too much in Q1, too little in Q2.”
All Athlons now being made in Dresden are Thoroughbreds. How many Thoroughbreds do you see out there now? Just 2200+, and not even too many places have them. AMD said they didn’t expect to be selling all Thoroughbreds until the fourth quarter.
Fresh product my butt. They’re waiting for the channels to work off their “stale” Palomino inventory while the “fresh” Throughbreds stack up in Dresden.
In 2001, AMD had a bit more than 20% market share. For 2002, the year-end figure will look more like 16-17%. We’re now talking about a company that will start selling Hammer next year with a market share that by then will likely be little if at all better than when they introduced the Athlon (that was 15%).
To add to AMD’s woes, bargain-basement Durons are continuing to increase their proportion of AMD sales, as they have been the past few quarters. It’s now probably pretty close to 50-50.
To break even (presuming flash memory continues to do OK, which it seems to be doing), AMD needs to make about $625-$650 million in revenues from its CPUs. They made $380 million last quarter.
It is very hard to see how AMD can get back up to those levels for the remainder of the year. To do that, they need to sell a lot more processors (this is conceivable for the Christmas quarter, though Intel should be in good production shape) for a lot more money (this is not, given the lack of new products and no shortage of Intel products). Indeed, it is easier to see them getting somewhat worse.
The AMD execs seem to agree with this assessment. Analysts kept asking variations of “When do you expect to start making money again?” and the execs certainly did not give the slightest hint that it would be anytime in 2002.
Their answer was more like “when we get Hammer, it will be a whole new ballgame.” They did say they didn’t expect to sell more Hammers than Athlons in 1Q 2003 (which no one outside of the dopey analyst who asked could have reasonably expected). What I found far more interesting was that they didn’t volunteer an estimate of when they expected that crossover to happen.
As I mentioned earlier, Intel provides very few new product details during its conference calls. AMD, on the other hand, tends to sneak bad news into them.
The major sneaked news at this conference call was Hector Ruiz saying that AMD would introduce 2400+ and 2600+ processors in the second half of the year. Please note what he did not mention; a 2800+ processor, which has shown up on the roadmaps in 4Q as the first Barton.
This is quite understandable because the only place that was supposed to make Barton (UMC) won’t be making any processors for production until 2003.
Hate to say we told you so, but we told you so about two months ago.
Very little was said about Hammer the product other than it was wonderful; no comments as to specifications. A few will get shipped at the very end of 2002 (almost certainly to OEMs), but the product announcement won’t occur until 2003 (figure a few days before Intel makes its traditional mid-January announcements, so let’s say January 9 or 16, 2003).
Wasn’t There Any Good News For AMD?
No. There doesn’t have to be, you know. Asking that is sort of like somebody in 1941 after hearing about Pearl Harbor asking, “Now tell me the good news.”
In 1941, I suppose one could have said, “our aircraft carriers weren’t around to be bombed,” and in 2002, I could say, “AMD’s flash memory seems to be holding its own,” but that’s not really good news, that’s “it could have been worse.”
However, just like Pearl Harbor didn’t make America instantly surrender to Japan, the terrible AMD results doesn’t mean inevitable AMD defeat and surrender. It does mean trouble, especially when serious relief may not come for another year.
The Final Question
The last question asked in the conference call came from a gentleman from Deutsche Bank. Unlike most of the other analysts, he did not ask creampuff questions. He first politely pointed out that the talk about market share was just a bunch of spin, then asked the question that really needed to be asked; the one we asked a couple weeks ago.
What he said boiled down to, “Your balance sheet deteriorated quite a bit this quarter: fewer assets, more debt. It isn’t going to get any better the next few quarters. Are you guys going to be OK?”
The answer he got was not all too convincing. Remember, what gets said at these things can get the SEC on your back if it strays too far from the truth. The AMD answer that came back was essentially “We feel confident we have enough cash to work through this period,” which in legal terms is not quite the same as saying, “We’re sure we’ll be fine.”
This hardly means AMD is doomed. It does mean there’s some reasonable basis for doubt as to the company’s continued existence if matters don’t improve considerably in 2003, and, of course, there’s some good reason to think that they will.
The reality is AMD is betting the company on Hammer, and they know it. They’ve even neglected their current product line to pull this one off. If they do, what happened last quarter and will happen for the rest of 2002 is not too big a deal.
But if they don’t. . . .
P.S. For The AMDroids Some will say, “This article wasn’t balanced and impartial!”
Intel Architecture (that including CPU processing) had a lousy quarter. “Lousy” is making almost $1.4 billion rather than $1.8 billion like they did last quarter. They made about 8% fewer processors, and got maybe $160 rather than $165 a processor.
AMD’s CPU processing lost somewhere around $200 million. They made 25% fewer processors, and got $63 rather than $75 for each one. They keep getting the same results; they’re going to run out of money in about a year.
You’re right there’s something unbalanced and partial about this. It’s called reality.