Originally created 04/30/01

Intel trying to stay ahead of 'Moore's Law'



LIVERMORE, Calif. -- Computing power hasn't doubled every 12 to 18 months just because Gordon Moore said it would back in 1965.

Few people understand that more than Intel Corp. chief executive Craig Barrett, who leads the company Moore co-founded in 1968 and faces the challenge and costs of maintaining "Moore's Law" -- while keeping prices low and convincing customers they actually need more powerful processors.

The continued prosperity of the high-tech economy depends on faster, better, cheaper chips.

Intel recently announced progress on two fronts. It plans to start making chips on larger, 12-inch wafers, cutting costs by about 30 percent over the current 8-inch wafers. And a new technology that adds more features -- more processing punch for smaller spaces -- also was unveiled this month.

During a recent drive from Santa Clara to Livermore, Barrett spoke about the challenge of maintaining Moore's Law, Intel's approach to research and the future of computing.

* * * * * * * * * * * * * * * * The Associated Press: Are you surprised at how long the semiconductor has been able to sustain Moore's Law?

Craig Barrett: I think everyone has been amazed that Moore's Law has gone as long as it has. I know that in 1965, when Gordon postulated that law, he didn't think it would be sustained over an extended period of time. That was 35 years ago, and we're still pretty much on track.

Most people would suggest it's got somewhere between 10 and 20 years more. That would make about 50 years of doubling every 18 months. Mother Nature doesn't do many things of that sort. It is quite amazing to all involved that something could be increased geometrically for that extended period of time.

AP: Do you and Moore still discuss it?

Barrett: We always talk about how much longer these things can happen. We talk about how much bigger wafers can get. How much more expensive fabrication facilities can become.

AP: As more computing power is available, how do you see the personal computer evolving?

Barrett: I see the computer continuing to do very much the same things it has done in the past -- word processing, spreadsheets and act as a vehicle to access information and to communicate and also to extend its features and capabilities into natural data types -- to audio, digital video imaging, entertainment, animation, rich communication.

I think the image of the desktop PC, which is this rectangular, beige box, flat and on end under the desktop or on the desktop is going to disappear, mainly because houses will be networked and there will be wireless connectivity. You can put the guts of your PC anywhere. It may well be some place out of site.

AP: Intel is spending more than $4.3 billion on research this year. Where is that money going?

Barrett: It's not necessarily to fund all aspects of fundamental research associated with computing and communications but to keep our eyes open at those institutions that are doing basic, fundamental research and at the appropriate time get involved as it approaches the stage of commercialization.

AP: In the past, some of the greatest inventions of the computer age have originated in corporate labs performing basic research. Why doesn't Intel follow that model?

Barrett: There are historic reasons, which is the founders of the company came out of Fairchild (Semiconductor). Fairchild had what we call an esoteric R&D laboratory that did not interface terribly closely with the rest of the company. It had a very difficult time transferring its research results into the mainstream of the company.

(Intel founders) Bob Noyce, Gordon Moore and Andy Grove at that time noticed they could do much better if they dedicated their research more toward the development and commercialization of technology as opposed to the basic creation of technology.

AP: What are the challenges of relying on others to do the basic research?

Barrett: If I see a troubling trend going forward in the United States, it is the shift of research moneys away from the physical sciences and toward the life sciences.

If we are counting on these universities to be the research laboratory for our industry and other physical science-based industries, then the slowdown of spending of R&D dollars has to be a troubling consequence.

I basically look at our industry as being a value creator. We create jobs. We create products. We create things that the United States can export. We create value.

I don't want to demean the life sciences or medical research, but in terms of value creation, to a certain extent, the health industry is more of a tax on the economy than a value creator for the economy.

AP: Shouldn't industry pick up some of the slack?

Barrett: It's the role of industry to participate, but I think it's the role of the federal government to be concerned about the long-term economic viability of the country, the technology base of the country.

AP: A lot of time, effort and money is going into faster processors. Yet many consumers say there's no need for such speed. How do you respond?

Barrett: I heard it when we transformed the 286 to the 386 the 386 to 486 the 486 to the Pentium the Pentium to the Pentium II, Pentium III to the Pentium 4. That has been a common comment through every transition we have made.

Unfortunately, we tend to always look at the next generation of microprocessor in light of our past experience rather than thinking about what we can do in the future.

AP: The dot-coms and the Internet seemed to offer new uses for more powerful computers. What happened there?

Barrett: Business models. Anybody who had a halfway smart idea could create an IPO and go out and be valued in the billions of dollars with no products, no revenue, no profits, no customer service, no nothing.

It was an unsustainable model, so we crashed back to Earth. The dot-coms had farther to fall, and they're squished flat.

AP: Is forward progress in technology at risk because of the power crisis?

Barrett: I don't think it will slow down the overall pace of technology. If anything, it will reawaken the debate in the United States as to what a successful energy strategy is. I don't think you could have that debate without reopening the issue of nuclear power.

Unfortunately, California has been at the leading edge of not doing anything -- don't build any new power plants, don't build any nuclear plants (with the belief that) it will take care of itself.

It's very clear it's not taking care of itself. There's going have to be proactive dialogue and discussion on how we get a rationale energy policy in place. California doesn't have one.