You read comments some allegedly informed members of the mass media have made about the CPU companies, you’d think they’d
be breaking into “Happy Days Are Here Again.” The real news is that Intel’s going to have a horrible 2001, and AMD a mediocre one.
In the fourth quarter of 2000, Intel processors powered 81.5 percent of PCs while AMD had 17.1 percent. In the first months of 2001, Intel’s share fell to 77.3 percent and AMD’s grew to 21.1 percent.
From Jerry Sanders (actually from paraphrased or better conference call transcript): “We continue to gain market share, now at 21.5%. . . . Long standing goal of 30% unit share by end of 2001. Plans for capacity for that. With current market conditions, we don’t think we’ll get their by year end, but we’ll make progress throughout.”
From Intel’s Earnings Press Release: “Gross margin percentage in the second quarter of 2001 is expected to be 49 percent, plus or minus a couple of points, down from 51.7 percent in the first quarter. Intel’s gross margin expectation for the full-year 2001 is 50 percent, plus or minus a few points.”
From this, we can derive a number of important points:
No matter how stupid Intel gets, it will continue to rule for the foreseeable future: They have the manufacturing capacity, AMD doesn’t, and won’t. Even after Dresden is running full-blast, we’re still looking at AMD not
being capable of taking more than a 30% market share. That’s AMD saying that.
Now this may not be your impression as to how things are going at all, and there’s a reason for that. AMD’s market is heavily skewed towards the consumer market, and if you’re a do-it-yourselfer, it’s probably even more skewed now in AMD’s favor. But a computer show is not typical of the whole industry. Intel still dominates big time in business computers. Sure, AMD may be able to crack that market a little more over the next year or two, but at some point, without new manufacturing capacity, it’s either supply the business end or the consumer end. This is going to be another reason (or at least excuse) for businesses to stay with Intel; businesses prefer to have reliable sources of supply than a somewhat better,
but not quite as reliable competitor.
For AMD to be even capable of making more chips than Intel, they need two more big fabs like Dresden. Looks like they’re going to build another one, but that won’t be ready until 2004 or thereabouts. They could possibly supplant Intel with one more big fab and contract out some fab work to places like TSMC or IBM, but those guys usually charge a pretty penny. At this stage, AMD shouldn’t and isn’t contemplating that. That might make sense in a few years under the right circumstances, but AMD is still building.
Intel’s going to have a bad year profit-wise due to Willy If you look at the numbers in the financial statement, you see something pretty amazing: Intel sold fewer processors, but it cost them more money to make them. Revenues went down one and a half billion dollars, but their costs went up $400 million.
This is due, of course, to the die size of the current Willy. It chews up a lot of silicon for its price. That means Intel can make far fewer Willys with their equipment than they could PIIIs. That hurts Intel’s profit margin.
Intel’s gross profit margin (gross profit is what you get paid for the product minus the direct costs of making it; gross profit margin is the gross profit divided by the revenues) dropped pretty dramatically. Intel’s profit margin has historically hovered around 60%. This quarter dropped to about 50% Not that you need to pass the hat around for them, and it’s still better than AMD’s slightly-over-40%, but that’s the main reason for profits dropping so much. What’s important to note is that Intel essentially said it wasn’t going to get any better during 2001. Indeed, it is likely to be tough to even maintain it. Intel is apparently going to be pushing those silicon-greedy Willys heavy for the rest of the year, and is slowly inching its way towards a possible price war. Well, maybe not a price war from your standpoint, but when Intel gets forced to charge just $350 for their best standard processor, that’s certainly a price war in Intel’s eyes.
Will it turn into a real price war? Good question. In the past, Intel’s been able to prosper when their machines cost $200-300 more than AMD’s; and that’s where the late April price cuts should put them. Will that continue to hold true? If it does, there will be no real price war. If it doesn’t hold true, and Willys still don’t sell, Intel has two
more tricks to play in the fall, and if they don’t work, then I think they panic.
Moving over to the AMD side:
This will be a lackluster year for AMD: Now lackluster isn’t too bad under current conditions, but even if times were (or turn out to be) better, it still wouldn’t be a stunning year. AMD made a lot of money out of flash memory, that market is now weakening. They’ll be able to expand Dresden and get production up somewhat, but again, nothing like 2000. In all likelihood, they’ll be able to take a few more percentage points market share out of Intel’s hide, but it will probably be tough to get much above 25%. AMD’s also going to be paying taxes like the big boys, too; practically all the drop in net profits from last year came from taxes AMD now has to pay.
This is a year of consolidation for AMD. They have themselves pretty well positioned to ride out any overall recessionary storm, provided Intel doesn’t panic and keeps those factoids mentioned above in mind.
Will Intel panic? I think the odds on that are greater than I would have thought a few months ago. Pricing a 1.7Ghz Willy at $350 is a sign of potential panic. Not all-out panic, but definitely higher-anxiety.
I suspect what Intel is thinking is “we’ll keep the price low over the next few months, see if that works. If it does, then when we shift over to DDR boards for Willy (and this is looking likely for the fall), we can put prices back up again. If not, we get more of a cost savings from DDR, and that definitely should throw those impudent AMDers back off the beaches, especially with .13 micron Willys showing up, and AMD being behind us there.”
This could well work. If it doesn’t, then we have a real good chance of scorched-earth in 2002.