In Hannover on Monday, AMD's chairman and chief executive, Jerry Sanders, discussed his company's future processor strategy, its continuing battles with Intel and his dream of seeing his processors in the corporate environment. ZDNet's editor, Richard Barry, was there.
Of all the underdogs in this industry, AMD must rank as one of the favourites as it wrestles what Jerry Sanders calls the "800lb gorilla" that is Intel. In a frank interview on Monday morning, the 62-year-old leader of the Sunnyvale processor manufacturer spoke openly about the challenges AMD faces as it prepares to take on Intel on three fronts: value, desktop and corporate.
But first, that comment from Michael Dell.
You've spoken about your dream of entering the workstation/server arena with your Sledgehammer project, but doesn't it worry you that a company like Dell is concerned about your existing technology, let alone a brand new one?
No I don't think so. It turns out that the success of the Athlon processor speaks for itself. Nine of the top 10 manufacturers use Athlon today. I think Michael's remarks were really not about compatiblity.
I was careful to ask Michael if his concerns were about compatibility, and he was very clear they were.
Then I would say there are no issues of compatibility. The issue that Michael was talking about, as far as I could piece together, was that there is not yet a platform for the commercial space that has all the requirements needed. We're the first ones to admit that. We're at the early days. For example, we don't have a two-processor solution yet. We don't have drivers for every imaginable video card or every interface card, and so on.
At this point in time, we don't have enough to offer Michael Dell's company for him to give up his preferred position with Intel. But we're not giving up on Dell, and I think, in time, it will be in Dell's best interests to see the value of competition. We're going to do just fine.
There are those within the analyst community that say the PC-centric era is over with the likes of PlayStation2 on the horizon. Do you conform to that notion?
It depends on whether or not you believe the living room is going to control the Web or the office. We believe the greatest opportunity will be the office using business applications on the Internet. Business applications will not be run on PlayStations. There's a better opportunity for consumer applications, and we do have plans, but we can't reveal what they are at this time.
Could those plans involve Microsoft's X-Box?
I can't comment at this time, but I would observe that the X-Box is a very impressive competitive offering to PlayStation2.
So you're excited by the X-Box?
I have no further comment. Really!
Tell us a little about last quarter and how AMD profited from Intel's problems.
First of all, we did not profit last quarter from Intel's problems -- Intel couldn't match our performance. They caused themselves problems. Our limitations were primarily infrastructure. You may recall there was an earthquake in Taiwan in September. We were developing a brand new infrastructure -- a new chipset and a new motherboard -- to support the Athlon. We were hand-to-mouth on infrastructure support for Athlon in Q4.
What Intel did, or didn't do, had no significant impact on our business. What Intel was unable to do, of course, was to match our clock speeds in volume with their Coppermine. Now that it is more aggressive on top speed, we believe that we will have a more competitive offering. We don't think Intel's problems had any material impact on our quarter.
What we do believe is that Intel's problems had a major beneficial impact on our customers' attitude to our future. Historically, Intel was considered to be invulnerable. I think that today it is considered to be nearly invulnerable. We're looking forward to the point when people start to think of Intel as vulnerable!
At this point in time, AMD is producing 850MHz Athlons in volume. Both Gateway and Compaq have announced they will have systems available for sale for immediate delivery. The bottom line is that we are producing high volume, leading-edge processors to the extent that we will continue to have a competitive offering. I think that our customers will wish us to stay in the game.
Can you tell us what the memory architecture will be for new Athlons?
Right now we believe Double Data Rate (DDR) is the way to go. That will be supported in the second half of the year. I would still point out that we have a licence with Rambus, and so we are certainly not averse to buying in a Rambus solution. We still believe the customers are the ones who should tell us what they want. The difference between AMD and Intel is that Intel tells the customers what to do. We think that is another reason we are going to have success in the market place, because we give the customer a chance to make money on this product, rather than being told what's available to them.
And going back to my earlier point, you may recall Intel had other problems -- like not having functioning DRAM because of engineering issues. This is unheard of for Intel. How can it have these problems all of a sudden? This is a company with 10 times our R&D spending, and 10 times our revenue. How could it have these problems?
I'll tell you it's because they've never had any competition before. When you have no competition, you issue things when you're ready to issue them. When you have competition, you have to come out quicker. Intel for the first time is being forced to return-serve. Up until now, AMD has been forced to return-serve. I don't think that it has much experience of this, and it has made a few mistakes.
But we never underestimate Intel, because it does everything it can to crush us. It's been trying to crush us since 1986. In the 1980s, there were 15 x86 providers worldwide, today there's only another one -- that's AMD.
Can we talk about Spitfire and the competition it's going to have with Joshua and the like?
We don't think that anybody other than Intel and IBM have the process technology to head up high performance clock speed devices. We do believe that VIA will have an opportunity with its offering in certain low-range applications, particularly in Asia. But the realities are that so far it has offered what I call 'P-rating' devices, as opposed to true clock speeds. With P-rating, you measure performance by whichever benchmark you choose against the Pentium running at that top speed. So P-rating a 500 means that on some benchmark, you're saying your device is as good as a Pentium is with 500MHz.
Our experience is that customers don't like that -- we went through all that with the K5. We know that IBM doesn't like that. We know that Compaq doesn't like that. We know the big manufacturers of PCs don't want to have their customers saying, 'I thought I bought a 500MHz part, when I really got a 400MHz part that would be the equivalent of a 500MHz part.'
Yeah, it's real confusing, so we abandoned that strategy. I hope you know that Athlon outperforms Pentium III on every benchmark, clock for clock. So a 600MHz Athlon outperforms a 600MHz Pentium III. But we don't give it a P-rating of say 650 or 700.
At the top end, where you're focusing much of your efforts with Mustang and Sledgehammer, do you expect a slow start with corporates because they're not familiar with your brand and technology in that arena?
No, I don't think so. I think it is pretty easy to get into the consumer market, because it is self-contained. The guy buys once: it works or it doesn't.
It is a much more complicated equation in commercial space. Intel is doing everything it can to keep us out of the commercial space. As you know, more PCs are bought for business applications than for consumer applications. And although AMD is able to get more than 50% of US retail market share in mobile, we've got something like 40% in desktop US retail.
You say, 'well wait a minute Jerry, you've got 50% of the mobile market, and you've got 40% of the commercial market, or the desktop market, how come you've only got 16% of the worldwide market share?' The answer is: because we don't have a very large market share of the commercial stakes, because up until now, we didn't have the ideal solution. Athlon is an ideal solution.
Now we have to have the chipset and the drivers -- the rest of the offering -- so that everyone is comfortable. We are real proud of this here, and I think you'll see some major breakthroughs by the end of the year.
Is your stated plan still to achieve 30% of chips at market share worldwide?
We want 30% worldwide market share of processors by the end of 2001. That's been our stated position since 1996.
So you're going to be fighting on three fronts -- low-end with the Spitfire, Athlon in the middle and the Sledgehammer and Mustang at the top. Is that not too many fronts for AMD to be fighting on?
Well... we'll see.
Well analysts, such as Ashok Kumar, believe you're 'back in stride', but reckon you may be biting off more than you can chew.
If we were trying to be Intel, we would probably have a difficult time. But we're not. We're trying to be the rallying point for the industry alternative to an Intel monopoly.
So we have people developing chipsets for the server... like Alpha Processors and others. Other people are developing chipsets for Web servers of a more powerful nature. We've got people doing two-processor server chipsets and eight-processor server chipsets. We have a number of people doing mobile chipsets and a number of people doing desktop chipsets. We've got 12 guys developing motherboards for these various applications. AMD is the rallying point for the industry alternative to the monopoly.
We believe that in the quest to offer competition in the market place, our partners will enable us to have successful penetration in all those things.