As Intel Corp. prepares to ship 90-nanometer microprocessors by years end, Intel Corp. CEO Craig Barrett is already looking ahead to 2004 and an era of increased IT spending. Buoyed by a generally favorable third quarter, Barrett assessed the progress of his companys Itanium processor thus far and explained Intels long-term strategy to eWEEK Executive Editor Stan Gibson and Editor in Chief Eric Lundquist at this weeks Gartner Symposium/IT Expo in Orlando, Fla.
Are you satisfied with Itaniums progress?
Im getting happier all the time.
Do you wish you were happier?
No matter which product you ask me about, I wish I were happier. Thats a positive comment about Itanium, in terms of adoption and software solution support, which have always been critical for the product. What were able to do from a raw performance standpoint is effectively second to none. If you look at price/performance, its clearly the winner. If you look at the movement of the product from high-performance computing to enterprise systems, were happy with the progress were making. The trend is positive.
With its transitional strategy to 64-bit computing, AMD [Advanced Micro Devices Inc.] seems to have captured a certain amount of mindshare that has heretofore eluded it. Do you look over your shoulder at all at AMD?
Only the paranoid survive. We have taken competition with AMD seriously for the last 25 years, and nothing has changed. And by the way, back when we were seeing who could introduce a gigahertz [processor] first, you were giving them just as much coverage, so this is not the first time they have had mindshare. There is a very healthy competitive environment between the two of us.
The [AMD and Intel] products are significantly different. Itanium is an enterprise product. I dont think that even AMD would say that the Athlon 64 Opteron is an enterprise-level product. I think they would position it in the entry-level server or workstation market. So I dont see that as clear competition. Maybe there is competition in high-performance multiprocessor systems. However, Itanium does a very good job there against anything thats in the market.
So, are you leaving a niche to AMD?
It depends on what the marketplace wants and needs in that space. We are not flying deaf and dumb. We do listen to the marketplace.
We have a 32-bit offering with hyperthreading thats very competitive in the marketplace. Whatever AMD is producing has some segment of the market. Theyll come in and undersell us by 25 or 50 percent, so they capture a part of a very price sensitive marketplace. Are we leaving that exposed? I guess we are.
There is not a desktop OS that really takes advantage of that 64-bit stuff today. If the marketplace expresses a demand for that then well look at it carefully and decide what our next move is. We havent been asleep at the switch on this. Our strategy is to meet what the market needs.
The competition with AMD does get testy at times. Hector Ruiz has used the term “bribing” to describe your wooing of OEMs.
Lets be very critical here. The conversation and competition gets testy from their side. I dont think you ever hear Intel executives or employees badmouthing AMD or accusing them of malicious intent or unfair practices of any sort.
If Hector has anything more than accusations, there are appropriate channels to which he can take his evidence supporting those accusations. The FTC has looked studiously at Intel a couple of times and has said it has found nothing that we do violates the ethics or laws of our business.
Next page: Intel lets products speak for themselves.
Page Two
The Microprocessor Forum was held recently, but Intel was not there. Any reason?
I guess we didnt have anything to say. We let our products speak for themselves. We have a couple of new products that were anticipating to be in volume production this quarter. For desktops and laptops [Prescott and Dothan] on the 90-nanometer technology. We let those speak for themselves.
Weve been introducing follow-on versions of the Itanium family, and we let that speak for itself. We have our own forum, the Intel Developers Forum, and we increasingly use that forum as opposed to the Microprocessor Forum.
Do you have any elaboration on Intels recent quarterly results, which were on the whole pretty good?
Wed like to see more strength in the U.S. enterprise market because that has ramifications with regard to U.S. competitiveness.
Gartners CEO Michael Fleisher says that 2004 will be different from 2003, when strategic investment starts to replace cost reduction. Do you agree with that?
Yes—for two reasons. The U.S. enterprise has not been investing significantly in the last several years, so there is probably pent-up upgrade demand that will stimulate the market.
Second, all this discussion of productivity, competitiveness and outsourcing will increasingly drive U.S. companies to recognize that to be competitive in the world, they will have to invest in IT infrastructure. Those two things will probably pivot IT spending up in the enterprise space.
Convergence was a big theme during your presentation. If Im an IT person, how should I think of this in concrete terms? If budgets are going to increase in 2004, how might that money be spent on converged applications?
Youd look at more mobile and wireless applications, particularly, the increased combination of laptop and handheld applications working in parallel.
At Intel, technicians in our factory use wireless handheld devices to speed up the process of entering data. And people like myself who are on the road or in the conference room can move from spot to spot with wireless connectivity.
Is there any danger that this move to greater wireless usage could be ahead of its time, like videoconferencing was, despite Intels advocacy?
The real difference between videoconferencing and wireless is that wireless is growing up organically, but videoconferencing was always a top-down push, in which executives were telling people they needed it. Before we introduced Centrino, you could go buy your PCMCIA card and stick it in your computer. Now, Im hearing from people who want to provide Wi-Max to rural and suburban communities that are never going to get broadband any other way.
Does the Bush administration ask for your advice?
I give it to them whether they ask for it or not.
Do they listen?
I think they listen as well as any other government. They are always preoccupied by re-election campaigns and polls. We routinely speak out on the issues we think are important. They are education, infrastructure, and research and development spending. They are policies over which the government has jurisdiction which we think either promote or impair competitiveness. [The move to expense] stock options [is] a great example.
Page Two
There is increasing concern that the technology industry is moving overseas. Your thoughts?
Taiwan, with a population of 25 million, makes 50 percent of computer peripherals. Lets imagine a country like China with two and a half billion people coming into the system. Would it be an impact 100 times as big? Even if its only 10 times as big, its cataclysmic.
Classical economists will tell you its wonderful; a rising tide lifts all boats: If they get the jobs, the standard of living and consumption power goes up and then well sell to them. But the devil is in the details. If they take all the jobs before they start buying things, is that a challenge? The one clear thing is you have to move your contribution to a higher level on a continuing basis.
Thats a source of consternation. When people see engineering and white-collar jobs going away, people say, whats the higher level, whats the next step up? There are no simple answers to that. Thats why education, R&D and infrastructure are so important. Those are the things that drive your competitiveness. If you dont do a good job on those, then you get left behind.
Thats the national debate Id like to see. But the debate were having is how to protect textile jobs in North Carolina. Were looking at the old equation.
Intel has become the platform of choice for Windows and Linux. How does that change your responsibility for security?
We, as well as the OS, the middleware and the apps guys, all have responsibility in that area. And the technology were coming out with, whether its Lagrande, virtualization or Vanderpool technology, is designed to provide some hardware assist to the hardware-software stack.
Were investing even outside of our domain, for example, in content protection, whether its broadcast flags or closing the analog hole or things of that sort. So, yes, we have a responsibility in those areas.
Looking ahead for the next decade, you have plans to enhance your microprocessors with multiple core technologies. Looking farther ahead, theres talk of carbon nanotubes and the like. How do you see the future?
The architectural innovations, the multiple core technology, are clearly going to happen. There is also tunable radio technology that is clearly going to happen, which will bring about seamless connectivity regardless of the protocol that you connect to.
Youre also going to see all sorts of hype about nanotech and thin polymer films for memory storage applications and some things. But if you want to look at the workhorse for logic, its the seamless transistor. Weve already got four or five generations working in our laboratories. Were introducing 90-nanometer this year. Weve already got the 65-, 45-, the 25- and the 15-nanometer transistors in our laboratories. That extension is predictable, and as long as its predictable, it is increasingly difficult for a disruptive technology to come in with the same economic efficiency that the seamless transistor has.
So theres not a window for a disruptive technology between now and 2010?
Its beyond 2010. It is clear that the seamless transistor will be replaced just as the vacuum tube was replaced. However, Im comfortable that the seamless transistor will be good for the next 10 to 20 years because weve already got the next five generations teed up.
Discuss this in the eWEEK forum.