and Rick Merritt
Intel Corp. chairman Craig Barrett sat down with EE Times to look back on his tenure as the company's CEO and forward at the future of the industry.
EE Times: So, consider this your exit interview.
Craig Barrett: I'm already gone.
EET: Well, what was good and what was bad while you were CEO?
Barrett: The typical Intel fashion is that we always talk about the bad stuff. The bad stuff is, we had a pretty serious recession hit the industry in the latter part of 2000 probably the biggest, deepest recession in the semiconductor industry in my 30 years in it. And it's never fun to live through one of those.
EET: Describe these eras in Intel history: The Noyce era, the Moore era, the Grove era and the Barrett era.
Barrett: Well, it's kind of difficult to differentiate between the Noyce and the Moore eras, really, because [Robert Noyce and Gordon Moore] were the two founders of the company, and they founded the company on the very simple premise of doing something with integrated-circuit technology and commercializing integrated circuits. I look at their era as heavily focused on advancing technology and then on what fell out of that advancement, whether it was SRAMs, EPROMs, microprocessors, microcontrollers, DRAMs things just continued to come out of the pipe.
Andy [Grove] took over when the PC started to build, and that's when he had the most influence on the company; he drove us hard in that direction, and we achieved a more reasonable market signature, market penetration, with the microprocessor. The IBM design early on helped, but we've had competition over the 25 years or so since then.
The last seven years have been an attempt to take the company to a somewhat broader direction. And so you've seen Intel invest a lot in communications technology in the last five or six years, and it will continue to invest in that area from this point forward.
EET: Can you give us kind of a report card of where you think we've gotten to in digital media vis--vis the content owners and the studios, and what's coming next?
Barrett: It's been a rocky marriage. We've always preached that there were a few things that had to happen in parallel to make a happy marriage and to bring digital content into the consumer's lap. Technology was part of that. Law enforcement was part of that, and reasonable commercial models that the end consumer would adopt. So there was always a variety of things that had to happen in parallel.
With digital music, clearly, there was never a consumer [business] model until Apple and other companies started to bring them forward. But by then, most digital-music content was already in play; we kind of had an open-source model for free.
EET: Let's pretend I'm your shrink and you've just come in for your $300-an-hour visit and
Barrett: You're too expensive.
EET: $250, but that's as low as I'm going, and the subject of China comes up. What are your anxieties?
Barrett: The anxieties I have are not so much China-oriented as they are about what China represents. China is an emerging market. It has 1.3 billion people, has a pretty good educational infrastructure, has a government that's very aggressive about competing in the future. So what it represents is really the competition to the established markets as we know them today the U.S., Japan and Western Europe.
So I have the same worry about India, about Vietnam and Thailand and the Czech Republic, as I do about China: It's competition, and whether companies will have a competitive advantage in the marketplace because of the environment in which they operate.
EET: That's keeping in mind Harvard business professor Clayton Christensen's notion of disruptive technology and the fact that the Chinese have a decent 32-bit MPU architecture that they might get to be just good enough where they can disrupt the traditional PC and microprocessor model?
Barrett: People have been suggesting for 30 years that we're just good enough; the 25-MHz 486 was just good enough. With the convergence of technology we talked about earlier, I see a continued need for performance and capability with Moore's Law going forward. One gigahertz is not good enough compared with a 3- or 4-GHz part; it's not good enough compared with a multithread and multicore part. Clayton has some very interesting ideas, but just because they're in print with his name on them doesn't mean that we subscribe to them.
EET: But the semiconductor industry is schizophrenic on the subject. Just in the last few days, I picked statements from Intel out of the paper saying, in effect, "We'll save a billion dollars in taxes if we locate our next plant in some other country. . . . On the other hand, we want to have better-paid science and math teachers."
Barrett: We're not schizophrenic. First of all, things like corporate income taxes are effectively diminutive as part of the U.S. budget. Second, if you do the math, creating high-paid jobs is the benefit that you give a country wherever you create those jobs. High-paid jobs carry with them something called personal income tax. Personal income tax is the primary form of supporting the nation.
Anytime you go to look at where Intel puts a major facility in Oregon, Arizona, New Mexico wherever we go, we build these multibillion-dollar facilities. What do we always negotiate when we go there? We negotiate the lowering of property tax. Why? Property taxes are supposed to pay for the families of the people who work in those buildings.
I have a $3 billion facility on one side of the street, and I employ 2,000 people. The guy across the street has a credit card center that cost $20 million and employs 200 or 300 people. I pay my employees three times as much as they pay their employees. Is it a net benefit to the community if I don't pay property tax there? Any study will show you, yes, it's a net benefit to the community.
EET: So what is the answer for improving education in this country?
Barrett: I increasingly think the answer is competition. I'll get on my soapbox for a minute: If you look at what we've got in the U.S., we've got "no child left behind," right? Are we doing OK or not doing OK? Is this system broken or not?
We now have nine states and several school districts suing the federal government, saying they don't want to be measured. So even putting the measurement system in, we have a problem.
Second, do we have a meritocracy in paying good teachers good wages, or is it a time and grade system, a union-based system, where no matter how good a job you do, you only get paid on the basis of how many years you've been teaching? So you don't have a meritocracy. You don't want to measure [performance] and when you do measure it, when you live in Arizona, say, you measure yourself compared with New Mexico or California; you don't compare yourself with India or China or Russia. So the only way I think you fix this ultimately is you put a performance-based pay system in place and you open it to competition.
EET: You took over in '98 and drove the portion in communications, the diversification of the multiprocessor sense of the company. Intel's still trying to get a little traction in communications, but you put the stake in the ground. If we got together seven years from now and looked back on this interview, how would Intel be different?
Barrett: Let me go back to rephrase your question a little bit. What did we do in the seven years I was CEO that will reap Intel benefits seven years hence? We clearly invested in basic R&D and process technology, extreme-UV lithography and all sorts of good stuff that I think will allow us to continue to be at the front of the parade in Moore's Law and driving transistors. I think we made the investments in manufacturing, and we'll continue to reap those benefits to the next decade.
We've heard [Barrett's successor] Paul Otellini describe the platformization of Intel and how we're going to organize ourselves and how we are organized to bring other platforms into the market like Centrino technology. I think you'll see that a lot over the next 10 years.
EET: What in your opinion is Otellini's biggest challenge in the next couple of years?
Barrett: It's the same as Intel has always had: For Intel to be successful, we have to be and be perceived to be a high-tech, high-growth company.
EET: In the White House, when one president leaves, there's some sort of tradition where he usually leaves something in the desk or in the Oval Office a bobblehead doll or a book or something like that. Do you guys do anything like that?
Barrett: I thought what the Clinton administration did was to take all the W's off the typewriter. I thought I'd take the B's off.
EET: Who's after Otellini? Pat Gelsinger [senior vice president and general manager of the Digital Enterprise Group]?
Barrett: Just as we never introduce new products before their time, we never discuss new CEOs before their time.
EET: You mentioned the competition in education and industry, and I'd feel remiss if I didn't bring up one difficult area that's come up during your tenure: the antitrust investigations that you and Microsoft had to go through. Intel seemed to go through it much more successfully than Microsoft, not needing to fight it out in the courts. But tell us about what happened recently with the Japanese antitrust allegations.
Barrett: The Japanese Fair Trade Commission basically came and said, "We don't think you should do these few things," and we questioned the validity of their arguments, we questioned whether they were really interested in protecting the consumer or protecting the competition. Most European trade commissions and the U.S. FTC basically are focused on whether harm is done to the consumer, not whether harm is done to the competitor.
So the Japanese said, "Here are a few things we want you to change." We said, "We don't think we've done anything wrong. We don't think that we've done anything illegal. You want us to change a few things? Fine. As long as we can continue to service our customers in the fashion we've done in the past. We don't think the things you're requiring will impact our ability to provide high-quality product, high-quality service to our customers.
EET: So those practices were offering discounts if you used 10 percent or less of a competitor's product or extra discounts if they offered none of the competitor's product. Are those practices OK, and can they still be used elsewhere?
Barrett: We don't disclose the contracts we have with customers in the press. Those were the two very vocal things that were raised sound bites out of much deeper relationships with customers.
EET: The practices that they said were anticompetitive do you think they are?
Barrett: I don't think we were anticompetitive in our dealings with the Japanese customers. No. I think we provided good value to those customers, and I think we provided great products at a time when our products were much better than the competition. We have competition that any one of our customers can buy products from whether it's AMD or somebody else [AMD has sued Intel in the U.S. and Japan for alleged antitrust violations. Ed.]. We're not going to restrict anybody's choice. One of our biggest customers is Hewlett-Packard, who is AMD's biggest customer.
EET:: We write a lot about the changing dynamics in manufacturing in the semiconductor business how there are fewer and fewer IDMs, it's harder and harder for a single company outside of Intel to spend the capital necessary to remain in touch with the next generation, offer lots of products and have a fabulous business model. Do you buy that theory that it's becoming more and more difficult for companies without their own factories to have a future in this business?
Barrett: Each generation gets tougher, and therefore the challenge of integrating all the processing steps to get a product out the door gets more difficult and requires a lot of R&D investment. Our R&D bill goes up each year. So I buy the fact that the foundry model was not one that was predicated on R&D investment; it was predicated basically on low cost and low overhead.
Who are we most worried about in the future as the competitors? It's the few vertical guys who are continually willing to invest and value technology. Samsung is a company like that. I think Texas Instruments is a company like that.
Remember T.J. Rodgers' model [for Cypress Semiconductor Corp.]? He was never going to be a big, fat, bloated company; he was going to create hundred-million-dollar little companies within Cy-
press, and he was going to grow those. And when he said [roughly a decade ago] that Cypress was probably approaching a $500 million company I think they're still approaching a $500 million company.
T.J. made a basic mistake. He wrote a book [No Excuses Management].
EET: So you won't do that anytime soon?
Barrett: I wrote my textbook [The Principles of Engineering Materials, with William D. Nix and Alan S. Tetelman; Prentice Hall, 1973]. I'll leave the book writing to Andy.