The fourth comes forth.

To paraphrase Craig Barrett's predecessor, it's wise to stay paranoid.

Robert Noyce, Gordon Moore, Andy Grove, and now...Craig Barrett. In Intel's 30-year history, its first three leaders achieved Mount Rushmore-like status in the American semiconductor industry - if not in world technology. Noyce, an inventor of the integrated circuit, Moore, who set forth a commercial principle that bears his name, and Grove, the exponent of "high-output management" and the man who explained why only the "paranoid" survive, were present at Intel's founding. Grove (1997's Chief Executive of the Year) steered the company through several near disasters and succeeded in making it the leading maker of microprocessors - essentially the brains for most personal and other computers systems - in the world. Earlier this year, Grove handed the CEO baton to Barrett while staying on as chairman.

A man of quiet, almost professorial, manner, Barrett, 59, is a native San Franciscan who left a Stanford University associate professorship to join Intel in 1974 as a technology development manager. With several degrees in materials science, he is the author of more than 40 technical papers on the subject, as well as a textbook, Principles of Engineering Materials, which remains in use today in universities throughout the U.S.

Barrett's 24-year career at Intel has been anchored in product development and high-volume manufacturing. He became a vice president in 1984, executive vice president in 1990, and COO in 1993. Although new to his current job, observers note that the 6 foot 2 inch Gary Cooper-like figure, who favors Western string neckwear and enjoys riding horseback on his Idaho ranch, has already placed his sizable footprint on the company. "He's a scientist first, not a technologist," observes Cypress Semiconductor CEO T.J. Rodgers. "It was his scientific approach that pulled Intel's then-mediocre manufacturing process into the powerhouse it is today. Grove tried but couldn't do it. In frustration he turned to Barrett who did."

Just as Grove faced one of Intel's so-called strategic inflection points when he took control in 1986, Barrett is taking over at a time when the industry's tectonic plates appear to be shifting again. The semiconductor industry is in the third year of a slump, its worst ever. And it's not just the effects of currency devaluations in Asia. Sales of PCs are falling short of expectations. Manufacturers have been forced to cut back to adjust to slower growth and to create build-to-order schemes, putting the brakes on chip demand.

Intel, which ships more than three-fourths of the microprocessors used in PCs, faces competition from compatible manufacturers such as AMD and Cyrix, both of which have made inroads, although neither has proven able to keep up with Intel's technology on a sustained basis. Intel's share of unit shipments is expected to drop from 86 percent in 1997 to 81 percent in 1998. Its 29 percent decline in earnings and fiat revenues for the second quarter of this year did little to lift industry spirits. Observers who had been predicting worldwide chip sales of $200 billion by 2000 are now thinking the industry will be lucky to get back to a 1997 sales rate of $135 billion. The global slowdown has already caused the company to cut back 3,000 from its 67,000 workforce.

The emergence of the sub-$1,000 PC, where Intel has 45 percent of the market versus more than 99 percent for high-end PCs, has also taken the wind out of its sales. Intel's low-end chip, Celeron, which has received mixed reviews, is its late entry into the sub-$1,000 PC market and costs PC makers approximately $140 to $180 - far below the $500 that Intel normally commands for a new chip. Falling chip prices don't necessarily mean lower profit margins if manufacturers can lower costs faster through greater manufacturing efficiency - a key strength of Barrett's. Nonetheless, bigger challenges may be on the horizon. It was easier to convince customers to upgrade when software advancements drove new PC purchases. Today, more users believe that low-end machines are powerful enough. Also, semiconductors may be redirected by the emergence of a new class of products - new generation wireless computing devices, and telephones, as well as Internet and home appliances linked to networks that require specific low-cost chips. (Intel's share of the 98 million non-PC device units in 1997 was 9.6 percent compared with Hitachi's 24.3 percent.)

Few doubt Intel's command of the microprocessor market. With plans to spend $4.7 billion on new plant and equipment and almost $2.8 billion on R&D this year, the company will maintain its market leadership. But it is sobering to note that on the very day the U.S. Federal Trade Commission filed antitrust charges against the Santa Clara giant for "monopolistic practices," the company slashed prices on its desktop PC microprocessors by 12 percent to 32 percent. It's a decaffeinated monopoly that can't control prices.

Widely respected, Barrett is said to be the right man to lead the $25 billion giant through what his predecessor might describe as another strategic inflection point. Given the industry shifts, will Intel's historic strengths continue to prevail? "Intel has always been a company where marketing triumphs over technology," says Jonathan Joseph, semiconductor analyst at NationsBanc Montgomery Securities. "Seldom is Intel the first to introduce a new technology or process, but once it sees a good one, it seizes upon it and capitalizes more aggressively and more rapidly than others." Yet, some IT experts, such as Randy Fields, chairman of Park City Group, wonder whether Intel is in the position IBM was in the '60s in terms of having created a market dominance that is no longer sustainable. "Up to now the market has supported chip branding and the premium price charged, but with the sub-$1000 PC, it's not the chip being branded, but finally the PC maker. When you get down to a $400 PC by the end of next year, followed by a $299 PC after that, can Intel compete in that market? It takes them perilously close to the D-ram market they got out of years ago."

Barrett agrees that if the future really is headed towards $500 PCs, "it won't be much fun in the playhouse, period. But we don't think that's where the world is going." He portrays Intel competing aggressively at every level of a market that's segmenting somewhat like the automobile business, where PC versions of Fords and Chevy co-exist with Caddies and Mercedes. As he relates in the following interview conducted at the company's Santa Clara headquarters, Barrett sees Intel growing through the creation of new users and new uses. He intends to play in every corner of connected computing, beginning with Internet-based transaction processing services - where he plans to make Intel the Visa of e-commerce. Following in the footsteps of Noyce, Moore, and Grove, his goal "is not to let them down, as opposed to coining a new law or a new mantra for Intel to march behind."

- J.P. Donlon

GROWING PAINS 

What is your vision of the future of the PC industry?

We're as bullish as ever on the growth of computers. Our vision of the future is still dominated by connected computers - a billion connected computers in 10 years. That's a lot of opportunity on the productivity side.

We think that an appreciable portion of our business within the next couple of years will come from the higher end of the processing spectrum. And then the periphery stuff would be the networking, hardware, and then the application - whether it's electronic commerce or digital imaging.

Today's PC market seems to be driven by decreases in microprocessor costs as opposed to increases in power. What does that mean for Intel?

The major issue is that the market is segmented, with standard desktop entry levels at $1,000 and servers that sell for much more. Intel has had to respond to this segmentation. We used to have one size fits all. We just introduced it at the high end and then it trickled down. We now design products for every segment and they don't trickle from segment to segment. That's been a major internal change in the way we see the desktop computer industry and the way we structure product development and position.

Ultimately, the question is, "Does that $1,000 category extend the marketplace or does it cannibalize the $1,500 to $2,000 PC?" I don't think the jury is in on that yet. If the market stalls and the $1,000 PC cannibalizes the rest of the market, that's not going to be pretty. Nobody makes a lot of money in that space. But I hope and think that ultimately the sub-$1,000 PC segment will expand the market, just as low-priced cars expanded the automobile marketplace.

How do you view Intel's position in the various market segments?

Our competitors decided to go after the entry level space almost by default. They perhaps targeted that space before we were fully cognitive about the current segmentation. But when we recognized that we swung our resources around and said, "We're not going to just isolate ourselves at the high end of this business. We will play aggressively at every segment." We want to be able to sell processors in the $1,500 to $2,000 PC space and we also want to do it in the work station and server space. We want to be the main player in each of those spaces.

How has the value creation model on which you are attacking the future changed?

I don't see us straying far from our home turf, which is connected computers. I don't see us going off into the hula hoop business, the appliance business, or whatever. There are ways for us to get more dollars of silicon per footprint, which is kind of an extension of a previous business. There's a big server business in which Intel's played in at the low end, but is increasingly playing in at the intermediate and high end. Networking is a big opportunity for us to grow a relatively small networking business that we have.

Is the priority bandwidth over processing power?

The two go hand in hand. Businesses are not bandwidth starved. The home or the small business is. But that portion of the market is very concerned about bandwidth and so are we. The biggest stumbling block is that neither the cable guys nor the telephone guys looked upon high bandwidth in the home as new value creation for which people are willing to pay. Instead they were hell bent on trying to steal each other's business. We'd love to see the government tell the cable guys and the satellite guys to get their collective act together and be able to get big pipes into the home. Then, you start to get away from it being about band width back into wanting enough computing power to take care of the data.

Is this why penetration in the home is stuck around the 43 percent level?

Roughly half of homes have PCs, but if you look at the homes with young children, it's closer to 70 percent. The driving force for PCs is families with children. For the rest, it's a keyboard phobia issue, and that's not about bandwidth; it's about ease of use.

What are your expectations for the overall market?

We've had about a year and a half of flatness due to a mixture of currency crises, food chain shortening, and clearing out the pipeline of inventory. We think the second half of 1998 we will start to see some recovery.

We have a lot of complex things going on today. It's hard to tell what impact the $1,000 PC phenomenon is really having on the market as a whole because of the various markets - Korea, Thailand, Indonesia, etc. - that have collapsed.

CHIPS AHOY!

Fifty-six percent of your business is outside of the U.S.; what's your view of overseas opportunities?

There are three inherent growth markets: China, India, and - aside from currency crisis - Southeast Asia. Indonesia has slowed down and in Latin America the hot markets are Mexico, Colombia, Brazil, Argentina, and Chile.

The established markets are Japan, North America, and Western Europe and you'd like those to continue to grow. But these other countries are growing at 50 percent, so you want to establish a presence. We do that in sequential steps. We go into a new market with a sales team and publicize the brand, then follow that up with engineering resources and then with the more permanent presence of the manufacturing team.

Wherever I go in the world, leaders in government, industry, and education all recognize that the future is not in low labor rates or natural resources. The future is in the knowledge-based industry. And at the base is how fast you can get data. So if you don't have the computer skills, capability, and infrastructure, you're out of luck. Sometimes that's understood more outside of the U.S.

How has the Asian economic debacle affected the U.S. market?

For many years, Japan had a much higher growth rate than the rest of the world. That all changed with the economic crisis, and computer sales are no longer growing. One of the things we look at is how do we make sure the demand folks who are predominantly located in poorly performing economies - Japan and Korea - are making capital investments so that in the year 2000, 2001 we will have a repeat of what we saw in the early '90s.

What's your strategy in China?

China is the fastest growing PC market. My forecast is that by the end of this year or '99, China will be the second-largest PC market in the world - passing Japan. There's actually a fair amount of disposable income in the middle class in China. With the one child per family system they have you end up with two parents and four grandparents all focused on the future of that one child. PCs tend to rate over cars and such things in countries like that because it's the most important thing you can do for this child.

Our strategy there is to work with local companies like Legend and Great Wall, to bring the latest technology to the marketplace. We have fantastic name or brand recognition in China. We took a very proactive stance several years ago and it's paid off.

In China - in Beijing or Shanghai - you see these little six foot by eight foot computer stores with the Chinese menu in the front. You pick out processor, hard drive, memory, and whatever, and when you come back four hours later, your computer is built. We have a pretty good basic structure there in terms of dealing with the big established computer company as well as dealing with this lower tier kind of mom-and-pop shop.

In the U.S., we take for granted that our access to the Internet is through a PC, because it's affordable, which is not necessarily the case elsewhere.

The beauty about a PC is that it allows you to interact - to create content, play games, and do a bunch of things. But that doesn't mean that standalone game machines like Sega and Nintendo will go away. By the same token, it doesn't mean that you won't have Internet access on appliances other than PCs.

You may have something in your kitchen that allows you simple access - anything from a standalone deal to a wireless pad connected to a PC someplace else in the house. It depends on what level of computing or interactivity you're interested in.

One of the interesting aspects of the acquisition of Digital's Alpha factory is that we got [the ARM] processor, which goes into products with a limited amount of interactivity, such as the TV set-top box, something that is only going to cost a few hundred dollars. So we'll play in that market.

How will Merced, the next generation processor for high-end servers, change our lives? Why is this important?

It will ultimately give a seamless transition from smallest laptop computer to a desktop computer to a workstation to the biggest database you can attack. What is this going to mean for you as a user at home? Probably not a whole lot unless you want to have a super high powered workstation. But for computer networks, this is a very strong building block for more cost-effective computing. First of all, it's a 64-bit chip, but it runs 32-bit code or 64-bit code. It does a much better job at accessing big databases and punching big numbers. So from the server workstation standpoint it does wonders and allows Intel to move upscale in that direction where we haven't played.

It will replace or compete directly with existing precision architecture, such as HP's PA-RISC, Compaq's Alpha, and Silicon Graphics' MIPS. So it becomes a basic building block for any of the workstation computer server manufacturers - Digital, Sun Microsystems, Silicon Graphics, IBM, Hewlett Packard.

What is the purpose of Intel's Microcomputer Research Lab (MRL)?

MRL's purpose is to look at computing architectures. It's an extension of the fact that for years the PC industry went with established textbook computer models and basically just integrated more and more transistors. You could integrate more and more functionality of known computer architecture. Part of the self-sustaining model you have to have as a player in this industry is not just the next one or two generations but what happens three or four or five generations out. We need to completely establish ourselves in that space. So we set up MRL and its charter is to look at everything from architecture to compilers to applications 10 years out and give us insight into what sort of hardware we should be thinking about designing.

TOMORROW'S TECH

What development excites your personal imagination?

I get very excited about voice recognition and ease of use capability. That's a major issue because there are no keyboards for Eastern languages and you have hundreds of millions of users left out in the cold until you bring an entry mechanism that's something other than a keyboard.

You can go out and buy voice recognition software today that does pretty well once it's trained. But it gets better with more computer power, and with the next generation of processors you will see a heavy application of voice recognition as an alternative to keyboard.

As a guy who made his mark in material science, what do you think about the recent IBM announcement about silicon on insulator (SOI) being the wave of the future?

When I joined Intel in 1974, SOI was right around the corner. It has been well recognized that it has the potential to go eight times faster than gallium arsenite. The offsetting balance has always been cost.

What we tend to do from a technology standpoint is let our actions speak about where we are. Other people make press releases. Regardless of what Steve Jobs' cute little ads suggest, if you look at any standard bread basket of benchmarks, he loses. He finds one benchmark in some far distant corner; the G3 processor outperforms the Pentium II. It's easy to publish papers. We just kind of plod along, and we're the world leader in what we do.

What about something that is not silicon-based?

We keep our eyes open for replacement technologies. The reason I mentioned gallium arsenite before in response to the silicon insulator issue is there is so much cost effectiveness in the standard silicon technology that we use today. It is very difficult for anything to compete with basic silicon technology, just because - Moore's Law - we can still see a way of doubling the number of transistors every 18 months.

For anything else to compete with that it has to have some infrastructural base and gallium arsenite never got there just because the cost effective base wasn't there. There's a lot of life in the silicon.

How long will your dispute with the FTC carry on? Are you prepared for a 10-year IBM-like struggle?

We don't try to force people out of business. The FTC is saying, "You're some form of essential utility and you're not allowed to use your intellectual property in your best interest." We think we have the right to use our intellectual property to benefit Intel. Intel invested in it. Intel created it. So we have a fundamental difference of principle here and that's one of the reasons why there is not going to be any interest in settling because settling would be conceding that principle and that's dead wrong.

You mentioned internet-based transaction processing services as one of your growth targets.

It's the Visa equivalent in Internet commerce. For a small set-up fee, we will provide all of the software and hardware capability for a business to get into e-commerce and we will get paid on a transaction basis. The second half of this year we will start producing some revenue. Hopefully it will continue to grow.

What other service businesses are you exploring?

We've got another one that's the PC equivalent of triple A - the American Automobile Association - providing everything from diagnostics to data back up. And there are other areas that are fertile for exploration, such as security, which is absolutely critical for economic commerce.

You follow three high-profile predecesors; as Intel's fourth CEO, what will you be remembered for?

I came here at the ripe old age of 34 after being a professor for 10 years, and I felt like the old man on the totem pole. I was surrounded by a bunch of 20-somethings had been in the industry for almost 10 years. I knew the principles, but I hadn't played with any of the circuits. But I got a crash course and survived a few ups and downs in the industry, and I guess I did all right.

Following in the footsteps of the three previous leaders of this company is an awesome task. Frankly, Craig figures he can be a success if Intel gets back on something that approximates its historic route. Growth is the major issue and growth only comes from great new products. Intel is not going to grow by stealing market share. AMD's strategy is to steal market share.

We worry about working with the industry to set standards in areas like how to get digital content on the PCs and how to have wireless connectivity between these devices, because that's what allows this industry to continue to grow. That's why we sponsor the New York Music Festival's fair where they had 200 night club acts streaming audio video over the Internet. One of the beauties of the PC is that you go to Japan, India, China, you take your application with you and plug it in and the damn thing runs no matter where you are.

The benefit to Intel is more uses, more users for the PC. You can understand everything we do with that simple phrase. Everything we do is designed to create more users or more uses.