Advanced Micro Devices Inc. is riding high, maintaining the strong momentum it built up last year. In addition to rolling out its first 64-bit processors and scoring major OEM customer deals in 2003, the Sunnyvale, Calif., chip maker refurbished its memory business and made a couple of key partnerships.
The moves helped AMD string together its first two consecutive profitable quarters—fourth quarter 2003 and first quarter 2004—in several years. Company chairman and CEO Hector Ruiz last week sat down with Michael R. Zimmerman, executive editor of news, and Jeffrey Burt, senior editor, in New York to discuss these and other issues.
Opteron was released a year ago and got off to a tough start. Was there a single point during the year when you realized, “A-ha, now were going to get traction with this”?
Was there a trigger? Was there an “a-ha”? I would have to say that it was when I went around—this is about a year ago—I went around talking to a lot of customers, the actual users of this product, and wanted to know how theyd feel about an AMD-based product.
Part of the crowd treated me with politeness. But at the same time there was some truth, and what I was learning was that they were all very excited about what they thought an extension to the X86 architecture can do. And they were very much looking forward to evaluating it.
Theres a difference in how you say things, like, Yeah, sure, well be glad to take a look at it, as opposed to, We really think this could make a big difference. And, Were anxious to see it. And when I kind of saw that sparkle in these peoples voices and eyes, I really thought, Golly, this thing will really get some traction.
Was that after you rolled out Opteron?
It wasnt a month earlier or a month later.
As far as OEMs are concerned, looking forward, can we expect to see Dell [Inc.] involved with Opteron as well?
Well, you know, Dell is an interesting story [laughs]. We have a great relationship with Dell. As a matter of fact, Ive talked to Michael [Dell, co-chairman] sometimes more than Ive talked to some of our other OEMs that buy a lot of products of ours. But I believe that its just around the corner when they are really going to be surrounded.
Be surrounded by other OEMs with AMD.
By other OEMs with AMD products. I believe, I happen to know two in particular—I cant unfortunately tell you their names—but I happen to know two major, major users that basically said theyre not going to buy Dell products because they dont have an Opteron solution. I think more and more of that will occur. And Michael is a shrewd businessman because he will react to that sort of thing more than he will to cheaper prices and that sort of thing.
Next page: Dual-core products will “shock the hell out of everyone,” Ruiz says.
Shocking Everyone
What do you think of Michael Dell personally?
Hes an incredibly smart guy. Shrewd. Very personal. I find him delightful to talk to. And I have a good relationship with him, to the point that either he or I can just pick up the phone and talk to each other. Theres no secretary that screens you out and calls you back. None of that. Its really a good relationship.
But hes facing a very challenging thing. Hes got a commitment for $5 billion of product from a company [Intel Corp.]. How do you ever flip that? Thats pretty hard to do. And I can understand that.
But you know, Dell prides themselves on not being a leader in things of technology, they pride themselves on being a strong follower. But I never dreamed that they would be dead last following, I mean, theyre running at the very end. Theyre going to be at the caboose of the [Opteron] train. And its just kind of surprising to me that theyd wait that long. We still have hope and confidence that theyll see the light.
Any other OEMs that are right around the corner?
You know, I dont know if there are. Truthfully, I couldnt tell you or if wed even be allowed to tell you. But Id think were going to have a number of Asian OEMs be very officially on board here, if theyre not already.
And of course, Fujitsu, Siemens in Europe already. In places where names dont mean a lot to you, because theyre not well-known, but eventually theyll be well-known. In China, weve got major players fairly involved with the products.
Observers say AMD is on track with its first year of Opteron, but the real measure of success for the year coming up will be your ability to branch out into other form factors, like blades and four-way servers. Are you comfortable that you have enough processors out there now?
We have roadmap that when you look 12 months out, its pretty firm. You look 12 to 24 months, and its almost firm. And then you look beyond that, and its always subject to modifications of the market. When we look out to, say, the end of 2005, we are enabling customers to really create a tremendous breadth of product lines.
One of the most powerful things next year is going to be our dual-core product. To me, thats going to really shock the hell out of everyone, because its going to be hardware-compatible, infrastructure-compatible, pin-compatible. I mean, people that have a 2-P system can slap in a dual-core product and end up with a 4-P system for the price of a 2-P. Thats been the biggest drawback, everyone tells me. What keeps them from going from a 2-P to a 4-P system? Its price.
As far as the blade market goes, does AMD have to do anything special to get the Opteron down and into blades?
There are a number of things going on in the blade market, theres this IBM-Intel alliance thats trying to establish a standard that no ones paying attention to. So, thats sort of confusing things a little bit.
Then you have other players, and each of them have their own view of what a blade should look like … . So, were just picking those customers were already working with to say, How can we help enable you get to a blade Opteron as fast as possible?
What are they saying?
Frankly, some are ahead of others because theyve actually been thinking about it for a long time. Others are confused by this Intel-IBM standard setting that they want to do. But well see Opteron blades before the end of the year.
Before the end of this year?
Yes.
Next page: Improving margins on the memory side.
Improving Margins
Were talking about CPUs today, but you have to be pleased with the success of your memory business, especially after Q1. Your memory business outsold your computing business by about $50 million but brought in a lot less profit. Do you see that trend continuing, and do you want it to?
The thing were planning to do is to improve the margins on the memory side at a pretty good clip. We have some fairly aggressive goals to improve that. And if you look at our fourth quarter to the first-quarter improvements on those margins, that was a pretty significant jump. Even though its still off, were in the 20-ish gross margin, which is low. But we came from nothing. The market had completely collapsed.
So, we believe the combination of two things, the cost synergies of combining Fujitsu and AMD together [to form Spansion] are just now beginning to be realized. And theyll be realized very strongly in the second half of the year.
The other part is that technology transitions—were moving so rapidly on technology on flash that the cost benefit of that is humungous. So, we expect our margins to take significantly. But they will never be the same as processors. The business model is different. It takes only, roughly, 6 percent of sales in R&D to support flash. But it takes double that R&D to support a healthy microprocessor business … .
However, in terms of operating profit, we think its possible to have similar operating profits with just different ways of getting there.
One could speculate that your positive net income could continue through the year if you were able to pull that off.
Thats certainly the plan.
Is Wall Street receptive to that?
I dont think we have credibility yet to get them to give us credit for that yet. We have to do it a few quarters in a row.
The European Commission announced recently that it is investigating whether the EU governments favored Intel over competitors, which is illegal under public procurement rules. How pleased were you about that?
Well, Im glad theyre finally taking a look at that because this is an archaic [issue]; procurement rules have been in place for a long, long time. And they havent been revisited in a long time. And therefore, they need to be brought up to date. And I think it will be good for everybody to have competition for government procurement and not be restricted to one CPU supplier.
How different is a procurement rule from a product certification process that a corporation establishes? How is it different from a government?
The thing with government is that since taxpayers pay for theirs, and taxpayers are employed by industry, so long as they meet the specs of the criteria, they cannot keep a particular competitor from bidding on government contracts. So long as they meet the specs of the criteria.
I think the same thing in industry; its no different, except the industry does it a little step in between and says, Anybody that meets the following specs is eligible to be qualified as a supplier. It means if you meet them in a market-driven capitalist society, you cannot legally keep someone from being considered.
From being considered.
Thats right. Thats all you want. You want to be considered. Today, were not able in some countries in Europe to be considered because the … specs say it has to be an Intel-based machine.