Who has sold the most x86-64 server CPUs?
The answer, of course, is Intel.
“WHAT!!!???” you say (well, maybe some other four-letter words got into your answer, too). 🙂
Well, from the Intel Analyst forum the other, you have this pretty little picture (page 42, all credits to Intel):
The key words, of course, being, “1 M (that means million) Xeon processor EM64T shipped over 6 months. > 3X competition’s (that means Opteron) volume in 1/3 of the time.
That means Intel sold over 1,000,000 x86-64 Xeons in six months, while AMD sold maybe 350,000 Opterons in eighteen months. Since AMD did sell some Opterons those first twelve months, the real head-on-head count the last six months is probably Intel’s 1,000,000+ vs AMD’s 200-250,000.
So it isn’t a 3:1 Xeon/Opteron selling ratio now, it’s a 4:1, maybe as much as a 5:1 ratio.
These numbers mean that not only that AMD is not conquering Intel in the overall x86 server market; it isn’t even getting close in the x86-64 server market.
Of course, even losing 4:1 or 5:1 to Intel in x86-64 server sales is still quite an improvement for AMD, where the ratio a few years ago was more like 30:1.
Nonetheless, I would bet dollars to doughnuts you had no idea what the truth of the matter was. You probably thought Opterons were wiping the floor against a handful of Xeons, because that’s the impression the media has left you.
Well, this is sort of like the 2004 U.S. Presidential election. The media rather liked Kerry, but who won?
Please don’t write me telling me that Opteron is better than Xeon. I would tend to agree with you, but your opinion (or mine) doesn’t matter in this election. It may well be right, but it’s just as irrelevant.
The only opinions that count in this election are the opinions of those voting with their wallets, and up to now, they’re heavily voting Intel.
Yes, they’re voting a bit less heavily Intel than in the past, and that’s good news for AMD. Unlike Senator Kerry, AMD doesn’t have to win this election to have a winner.
But they aren’t winning this competition. It’s not even close.
The Underdog Bias
There’s a bias to the nonestablishment tech press, a geeky bias. It tends to side with the underdog, especially when the underdog is geekier than the overdog.
So there’s a bias towards AMD. There’s a bias towards Mozilla and Firefox. There’s a bias towards Linux.
If you say, “But the establishment tech press is biased towards the overdog, to Intel and Microsoft,” there more than a bit of truth to that, too. But two wrongs don’t make a right. Bias without rational cause is bad, period.
Unfortunately, that bias often so blinds the biased that they tend to report all the positives and little if any of the negatives.
Now it isn’t that AMD or Firefox or Linux don’t make products with at least some positive aspects to them that some people can reasonably prefer over Intel or IE or Windows, but too often the preference becomes ideological, a religion without a God. It becomes a faith.
This is why I might seem so anti-AMD sometimes; it’s a response against this bias. Nor does that make me pro-Intel; I’d say right now I’m probably more anti-Intel than anti-AMD.
Nonetheless, reality is reality. Companies like Intel and Microsoft are 800-pound gorillas in this business. You don’t have to buy from them, but you can’t wish them out of existence.
Like it or not, Intel outsells AMD in x86-64 server processors. If you were led to believe otherwise, you were misled (as well may have been those who directly led you), and that’s not good.