Let’s get these AMD production figures and claims into perspective.
This article has caused some discussion lately.
First, let’s point out what it did and did not say. It did say AMD processors sales went up 40% in 2001 over 20% to reach about 30,000,000 (which is just about right).
It did not say AMD processor sales would go up 40% in 2002 (though that’s what AMD seems to be planning, more on this later).
This 40% rise from 2000 to 2001 should not come as any stunning news. AMD has been pumping out processors at about a 30 million per year pace for most of the year; in the second and third quarter, they sold 7.7 million . Remember, also, that in 2000, AMD was still transitioning out of the K6 series, stopping for a while at Slot A, before finally settling in to socket A. Dresden was still being outfitted for most of this period.
So yes, AMD production has gone up 40% from 2000, but the year being compared had somewhat lower sales than might have otherwise been expected. Still a very healthy gain, though.
However, to get that healthy gain in units solds, AMD had to engage in very financially unhealthy price-slashing, dropping prices to previously unheard-of levels for leading processors to get the product off the shelves most of the year. That makes those sales figures less rosy than they might seem otherwise.
If you want to know how well a company is doing against its competitor, you have to know the size of the entire market, not just what the company did.
If the world market for a doodad is 100,000,000, and I make 1,000,000 in year one and 3,000,000 in year two, I’ve tripled my production. That’s good, but my market share only increased from 1% to 3%. If my other competitor has the rest, this by itself is not going to worry him too much.
During the K6-2 era, AMD was getting about 18% market share. It declined a bit during Athlon transition, then moved up to about a 22-23% market share plateau for most of 2001. So Athlon got AMD 5% more market share. Not bad, but not earth-shaking, either.
The Digitimes article indicates that AMD is positioned to grab about 27% market share in the last quarter. Presuming that’s correct, this will probably translate to something in the neighborhood of 8.5 million processors shipped by AMD for Christmas, or about a million more than they’ve been shipping lately. Certainly good, and a production record for them, but it’s only about a 15% increase from what they’ve been doing lately.
(The better news for AMD investors is that AMD looks like it is getting more money for its processor, very roughly $75 ASP compared to about $60 last quarter.)
The point of all this is that there has been no recent, sudden revolutionary change in the CPU market, no huge sudden gain for AMD. The world hasn’t suddenly discovered AMD the last couple months.
(BTW, 27% would not be a marketshare record for AMD. The best AMD’s ever done is 30%, way back when they had the 386DX-40, and Intel was transitioning to the 486. That went away pretty fast when the 486 came along.)
AMD is getting this 4% jump just as Intel has been announcing PIV shortages left and right. Combine that with AMD suddenly being able to get firmer pricing, and it’s silly to pretend that the shortages have nothing to do with this sudden blip. Not even the AMD representative in the DigiTimes article pretended that.
On the other hand, the size of the shortage factor shouldn’t be exaggerated, either. AMD was selling a solid 22-23% before this, so at most the Intel Shortage Factor is 4-5%, and it’s probably somewhat less than that.
Stop Being Avis, Start Being Pepsi
In the long run, what happens for a quarter or two has no big impact on the CPU market, just like AMD getting 30% of the market for a quarter back in the 386 days had no permanent impact.
However, AMD’s production plans presume yet another 40% jump, from about 8.5 million CPUs the end of 2001 to over 12 million by the end of 2002.
That all by itself is roughly 10% of world demand, and would put AMD well over 30% of anticipated world market. On top of that, it looks like AMD is negotiating for yet even more production capacity.
AMD wants to stop being Avis and start being Pepsi, and 2002 is the transition year.
Since the CPU market is hardly going to expand 40% in 2002 (10% will probably be more like it); further expansion is going to have to come out of Intel’s hide.
Just by going to DDR boards, Intel will have a more competitive product. As Northwoods become common, that will give Intel a further boost, and free up capacity now used to make humongous Willies to make more, smaller Northwoods.
This is not to say at all that Intel is going to blow AMD away, just that AMD will be facing a more formidable lineup.
The AMD execs already know and have said in their conference calls what’s going to happen. They’re going to start a price war, and think they can beat Intel with it because they can make their processors cheaper than Intel can.
You don’t harp on the difference between maybe $60 to make a .13 PIV vs. $45 to make an .13 Athlon if you think you’re going to get $150 for it.
It’s really hard to see how we don’t get back into a price war situation towards the end of next year if AMD follows through on their plans. Not unless CPU demand bounces back a whole lot more than anybody’s anticipating.
A lot more supply combined with shaky demand usually equals a lot lower pricing.