Eye2Eye with AMD's Jerry Sanders, Part II

Continuing ZDNet UK's interview with AMD's Jerry Sanders, editor Richard Barry questions the chairman and chief executive on his future, the push of the processor frontier and why he thinks Microsoft is 'God'

How important to AMD is support from Microsoft and, say, the Linux community?

It turns out that we clearly need support from Microsoft and Linux, and everyone else in the 64-bit arena. Our belief is that we have a compelling solution customers will want to adopt. The customers' demand will encourage the operating system support. I really can't say more than that, because to say more will cause some issues I'd rather not deal with right now.

(Sanders refused to be pushed on this point.)

Let me say this, though. When we said we were going to go to a 100MHz front-side bus with the K62, up until then the front-side bus was 66MHz. We caused the infrastructure to move to 100MHz. The second thing we said was we would come out with floating-point enhancements in the instruction set, which we called 3DNow!. In order to get 3DNow!, we had to get support from Microsoft. To my knowledge, it was the first time in history any x86 instruction set was supported by Microsoft that wasn't originated by Intel.

So, we're working with the operating people and the systems people, as you might expect. And we're encouraged by the response to date.

I guess everybody knows that if you run Linux, it runs great on Athlon. Unix runs great on Athlon. But that's not the issue. The wonderful thing about high technology is that you are always pushing the frontier. When you have a better idea, there is support for it.

But can you get the volume?

That's the one issue AMD has to deal with. So far you've been very kind: you could have said, "Goddamn, you're losing your shirt, you didn't have to for three years. Why is it going to be different in the future?" The reason we lost money in 1996, '97, '98 and '99 was because the industry was in a downturn, and we invested heavily in R&D and in the [fabrication] plant so that we could get the technology and volume to mount a legitimate challenge -- a credible challenge -- as an alternative to Intel's monopoly.

Our break-even levels are over $900m (£558m) a quarter. It's a pretty big piece of business, which you have to do to be in the game. With the success we've had, because of better ideas, our revenues last quarter were $970m (£601). We made some pretty damn good money.

So, I'm gonna admit that what we need is volume. And I believe we are going to get it. That volume will provide the infrastructure to support us. You won't have 12 motherboard guys supporting you if you don't make a lot of profits. We are going to produce 25 million processors this year, and with the grace of God, we are also going to sell them.

It's been rumoured that IBM has been working on a deal with AMD in the workstation arena. How close are we to an announcement on that?

We never comment on our customer's products before they are announced.

How close are we to the 900MHz Athlon announcement. You said it was "imminent".

Er... soon.

It's my mother's birthday next week, can I buy her one?

I'll sell you one right now! It'll be very soon. (Since this interview, ZDNet has learned AMD is trying to roll out its 900MHz, 950MHz and 1GHz processors simultaneously -- as early as this month.)

I'll explore other avenues on that one Jerry! How long does AMD intend to support the socket architecture?

It turns out it's an economic issue. We're delighted with the demand we've seen this quarter for what we call SuperSet -- that's Socket 7 running at 100MHz. The volume is much stronger than we thought it would be. As long as there is volume, we'll support it. But let's talk about Christmas 2000. What is going to be the entry-level point for Christmas 2000? In my opinion, it's going to be a 700MHz or faster part. And with that, you know a 100MHz bus is not a very exciting solution. A 700MHz part with a 200MHz bus is a lot more interesting. So take a look at Spitfire.

Intel has already demonstrated its 1.5GHz Willamette processor in what is seen as a bit of a game. Are you going to continue playing that game?

If the game means we can produce high-volume production of competitive performance processors like Intel, then the answer is yes. If the game is announcing or demonstrating a leading-edge product a year ahead of introduction, I don't know if I want to even play that game. Intel can put more engineers on marketing and promoting, but let's look at mid-year. Let's see which company has delivered more GHz-based machines -- AMD or Intel? Let's take a look at that, not who will, at some distant point in the future, have a 1.5GHz processor. We believe in doubling performance every 18 months. We can do that.

My understanding is that the Intel 1GHz machines, which were Intel-based PCs that were announced, semi-announced or pre-announced at the Intel Developer Forum, are in Intel boxes. There's no difference if it's company A, company B or company C -- it's an Intel-built box with an Intel power supply, Intel tweaking and Intel cooling. It's all Intel. All you do is put your brand on the box. That makes companies nothing more than distributors, and there's no margin in that. That's not the game AMD wants to play. We want to give the guy the chance to differentiate his box.

If your question is: do you think you can keep up with Intel on processor performance? Well, the answer is an unqualified 'yes'.

Micro Design Resources says one of biggest issues facing AMD in the coming two years is chipset support. It says you don't have enough offerings on the chipset front to compete with Intel, particularly as you've entered the high-end arena. What is your response to that?

I would say that Micro Design Resource is certainly correct that Intel has more resources to come up with more chipsets than we can. We're going to have the chipsets that we need to make a significant penetration in our chosen market. I don't expect that we would have the proliferation of chipsets Intel has. But I believe that the industry will be fine. We've got volume capabilities and a good idea. We'll get the chipset support, but yes, in reality, it is a concern. I mean, Intel is trying to exclude us by having many alternatives. We need to have the right solution for the target market, and I think we'll have that.

Rumour suggests that Intel makes life difficult for companies, particularly chipset and motherboard manufacturers, if they work with AMD. Can you comment on that?

Fear is a very strong factor in the market place. Did you see that TV commercial, with the thunder and lightning? The voiceover said: "don't mess with mother nature". It turns out there is a lot of fear that if people cooperate with AMD, Intel won't give them the kinds of preference they want. That is a real concern. I believe we'll work through that. I can't say it's behind us. The realities are, we are up against a very tough competitor that plays hardball every day.

That's all I can say.

But I'd just like to say as an aside, I don't think I could have persuaded the president of the $7bn or $8bn (£4.3bn to £4.5bn) Motorola semiconductor operation to come over and be my number two guy if he didn't think the prospect for Athlon and AMD were pretty good -- despite what the "gorilla" does. You don't leave a 25-year career as an officer of a corporation like Motorola, where you are the top guy in a $7bn (£4.3bn) business, to become the number two guy in a $3bn (£1.8bn) business if you don't think you've got a future.

Jerry, will Hector de Ruiz (formerly of Motorola) take over from you when your contract runs out at the end of next year?

The plan is for Hector to be my successor. My deal with Hector is that when I stand down as chief executive at the annual meeting in 2002, he will be my successor. And I hope that's the way it works out, because I sure as hell want to get somebody to pick up the time so that I can just be chairman.

So you can oversee the empire continuing?

I'd like to see it continuing to real profitability.

Would it be fair to say that AMD focuses too much on Intel?

First of all, we are not focused on Intel. I have to point that out. Intel is a very formidable competitor, but what we are focused on is the customer. 95% of PCs built today use the Microsoft operating system. Microsoft and AMD have a great relationship. We love Microsoft. We work very closely with Microsoft. Microsoft is God. OK? We have never competed with Microsoft. We do every thing we can to enhance and make Microsoft's operating systems even more widespread than they are today. It is unfortunate that the other guy who does this is Intel. I wish that there was a weaker competitor, but this is the hand we were dealt.

We didn't know we were dealt this hand back in the 1980s. We really thought we had a good relationship with Intel, and that we were going to coexist. That is, until they didn't honour their agreements. We had to fight a lawsuit for eight years, during which we suffered some real losses in momentum, and Intel went on to become the monopoly we know today.

But you know this is where the money is. There is no segment of the semiconductor industry as big as the $27bn (£16.7bn) processor business, which is going to be $42bn (£26bn) in 2002 or 2003. We're not to walk away from this $40bn opportunity. You know, people got all upset at the DSP market. The DSP market in 2003 is forecast to be $10bn to $12bn (£6.2bn to £7.4bn). Less than half the size of the processor business today.

I absolutely refuse to believe that there isn't room for two companies in a $40bn (£24.8bn) market. I mean, ah -- there is!

In part three, Sanders explains why he thinks Intel is abandoning the 32-bit market, and continues his views on the 64-bit arena. In the final instalment of this Eye2Eye interview, Barry asks Sanders about the rumours that have shaped the microprocessor industry and that spat with Gateway.

See also Eye2Eye with AMD's Jerry Sanders Part I.

See also Eye2Eye with AMD's Jerry Sanders Part IIII.