The New Intel

Craig Barrett is leading the chip giant into riskier terrain

Intel Chief Executive Craig R. Barrett sounds nearly poetic when he describes why it has been so darn hard getting the giant chipmaker to charge into new businesses--and into the Internet Age, where the old rules of computerdom no longer hold. Not surprisingly, Barrett conjures up a Western metaphor. He does, after all, live in Arizona, commuting most weeks to Intel Corp.'s Silicon Valley headquarters. Barrett compares Intel's microprocessor business to the creosote bush, a tall desert plant that drips poisonous oil, killing off all vegetation that tries to grow anywhere near it. Microprocessors so dominated the company's strategy, he says, that other businesses could not sprout around it. Chips, he says, "are a dream business, with wonderful margins and a wonderful market position. How could anything else compete here for resources and profitability?"

How, indeed, unless you have a CEO who is kicking up a sandstorm to find a way. Nearly two years after Barrett took the reins at Intel, the chip giant is in the midst of a historic overhaul that is transforming its business and its culture--for a second time. The first, back in 1985, Intel fled the memory chip business and bet the farm on microprocessors, turning itself from a diversified maker of chips into one focused solely on producing the electronic brains of personal computers. It was a brilliant move that set up the company for a golden period of growth under legendary CEO Andrew S. Grove. But now, the days of Intel concentrating virtually all its energy and investment on PC chips are gone.

The Grove era is over. Instead, Barrett is reshaping Intel into a supplier of all sorts of semiconductors for networking gear, information appliances, and, of course, PCs. More startling, he's taking Intel into radically different terrain, such as e-commerce, consumer electronics, Internet servers, and wireless phones. "We're putting a new image on top of the big powerful chip monster that eats the world," Barrett says.

And how. Last September, Intel unveiled a new family of chips for the networking and communications gear that zips data traffic through the arteries of the Internet. That's a $7 billion opportunity--and it's growing 30% faster than PC processors. In the same month, Barrett launched an even wilder scheme: Internet services, a $3 billion business worldwide that is nearly doubling each year. Intel opened the first of a dozen gigantic computer centers that it's building around the world to run Web and e-commerce sites for other companies. Over the next two years, the chipmaker will spend $1 billion on this elaborate network of Internet data facilities.

WEB WAR. Barrett's handiwork didn't stop there. In October, he paid $1.6 billion to buy DSP Communications Inc., a leading maker of digital wireless phone technology. Then, in January, Intel rolled out an ambitious plan to sell branded information appliances--screen phones, e-mail stations, TV set-top boxes--through phone companies and Internet service providers. And in February, the company barged into yet another new business, announcing a family of special-purpose network servers that manage Web traffic. The boxes will go head-to-head against gear from networking powerhouse Cisco Systems Inc. and a host of smaller rivals. "Craig stepped on the gas much more aggressively than I would have," concedes Grove, now Intel's chairman.

Each of these new schemes is ambitious in its own right. Taken together, they're a watershed. Within five years, Barrett intends all of Intel's new thrusts to be $1 billion-plus businesses and No. 1 or No. 2 in their markets. He's betting they will help Intel grow 15% to 20% a year, up from its paltry 8% compound growth over the past two years. So far, Wall Street is buying Barrett's vision, driving Intel shares up 40% since Jan. 1--easily the best growth among the 20 most widely held stocks in the U.S. By 2001, figures Merrill Lynch & Co. analyst Joseph A. Osha, products other than processors will make up a quarter of Intel's $38 billion in revenues and contribute nearly a third of its revenue growth. "Barrett is undertaking nothing less than a reinvention of Intel," says analyst Drew Peck of SG Cowen Securities Corp.

But no one is sure what he'll wind up with when he's done. Intel already is late to the Internet party. And the company is trying to break into new markets that already have scores of entrenched competitors. The result: Two years into Barrett's makeover, much of the payoff for Intel remains in the future. PC components are still the heart of its business, generating 90% of revenues and 100% of profits. "No organization its size can turn on a dime," says Peck. "This will be a slow, ponderous process, and meanwhile, expectations are very high."

Make that sky-high. Investors already are treating Barrett's plan as if success were a sure thing. That doesn't leave much margin for error. Barrett has to ensure that Intel's cash cow microprocessor business keeps throwing off beaucoup bucks to pay for the new gambits. In early March, for example, Intel will unveil the fastest PC chip ever sold, a Pentium III that runs at one gigahertz, or one billion cycles per second. But Intel's track record in its core business last year wasn't so good. The company had a rash of blunders in 1999--microprocessors and chipsets delivered months late, embarrassing design bugs, and supply shortages. Even loyal customers like Dell Computer Corp. and Gateway Inc. have taken the highly unusual step of publicly blaming the chip giant's gaffes for their recent earnings problems. Gateway was so incensed over supply problems that it's giving some orders to archrival Advanced Micro Devices Inc., which has caught up to Intel in chip performance.

That's prompting analysts to wonder if top management is prepared to handle the swirl of new initiatives. For starters, Intel is heading into territory unfamiliar to its executives--all of them veterans deeply rooted in chips. Analysts worry that the company's pell-mell rush into new businesses lacks focus. "They're throwing spaghetti against the wall to see what sticks," complains analyst Jonathan J. Joseph of Salomon Smith Barney. And rivals snort that Intel lacks key expertise in networking and data services--though they admit that its rich profits give it the means to buy into new markets. Barrett himself concedes that in its new endeavors, Intel will have to "compete, scratch, and claw for market share"--a bracing change from Intel's near-monopoly in PC processors.

PC POTHOLE. Barrett had little choice. After 10 years of 30%-plus compound annual growth, Intel hit a milewide pothole in 1998. Earlier attempts to expand into new businesses such as modems and video conferencing had gone nowhere. Then falling PC prices, computer industry consolidation, and increased competition piled on top of one another, causing Intel's revenue growth to slow to 5%, while earnings declined for the first time in a decade. The bad news drove Intel's stock down 30% and kept it off its peak for most of 1998.

The biggest culprit in Intel's slowdown was a changing PC landscape. For the first time since the IBM Personal Computer exploded onto the market in 1981, PCs were losing some of their luster. Instead of clamoring for more power to run fatter software programs, many customers just wanted cheap PCs to get online. Low-cost PCs meant low-cost chips, something rivals realized faster. Peering into the future, says Frank Gill, Intel's former top networking exec who retired in 1998, "Intel could see that the next 10 years wouldn't be as lucrative in processors." Adds Barrett: "It used to be that the PC was the center of the action, but now it's clearly the Internet."

There's no question that Intel is now a believer. Just look at the company's semiannual developers' conference on Feb. 15. As 3,200 digit heads crammed into a ballroom at the Palm Springs (Calif.) convention center, giant video screens flashed to life and the unmistakable riff of Steppenwolf's Born To Be Wild blared into the hall--only it had new lyrics: "Get your modem running, head out on the I-way. Looking for e-ventures, and whatever comes our way. Born to be wired." It was the warmup for a speech by Grove, the legendary chip warrior who barely sounded as though he were still in the semiconductor business, as he went on about e-commerce, gigabit networks, and facilities filled with servers. "For the first time in 15 years, we have found it necessary to change our corporate mission statement," said Grove. Now, instead of being just the leading purveyor of PC technology, Grove says, Intel aims to make the building blocks for the entire Net Economy.

To do that means undoing much of what Grove put in place. Within weeks of taking over in May, 1998, Barrett began dismantling Grove's rigidly centralized management structure, eventually breaking Intel into five groups whose managers report directly to him. He loosened Intel's conservative financial management to let more of its $12 billion cash hoard flow to acquisitions, equity investments, and internal startups. "We had to change the culture and the way we run our business," he says. Most of all, Barrett had to wean Intel from the defining strategy of the Grove era, when growth lay in stimulating demand for ever-more powerful PCs.

To flesh out his vision, Barrett has been on a spending spree. He's buying companies to beef up Intel's product line and help shed its notorious "not-invented-here" syndrome. In 1999 alone, the company spent $6 billion snapping up 12 companies--more, for the first time ever, than it spent on capital equipment, and nearly as much as the research-and-development and capital budgets combined. In a bid to mine fresh ideas within the company, the 60-year-old CEO has poured some $50 million into 25 homegrown startups that could someday become new product lines or be spun off as subsidiaries. And hundreds of other ideas are in the pipeline. Gerhard H. Parker, the Intel veteran who heads the company's New Business Group, says people are so excited about the opportunity that an employee recently chased him into the bathroom clutching a business plan. "It's wonderful to see that kind of enthusiasm," Parker says. "It wouldn't have happened a few years ago."

BATTING AVERAGE. Not without Barrett. While Grove is known for his fiery temper, Barrett is cool as ice. In his years before becoming CEO, he was known occasionally to bring a baseball bat to meetings to ensure--humorously--quick capitulation from intransigent colleagues. A native of the Bay Area, he studied metallurgy at Stanford University and went on to become an associate professor in the school's engineering department. After a sabbatical year working for Intel in 1973, he returned to Stanford for one day, then chucked it all and went back to Intel for good, rising through the ranks on the strength of his operational skills. He's credited with turning the company from an 80-pound weakling into an 800-pound gorilla in manufacturing, largely through inventing a technique called "Copy Exactly" in which every Intel plant is identical, down to the colors of paint. That made it easy for Intel to roll out new production techniques to all its factories--and dramatically boost quality and productivity.

So how did Mr. Manufacturing get Intel's troops jazzed about something besides their beloved microprocessor? He didn't need a baseball bat. When Compaq Computer Corp.'s $999 PC hit the market in February, 1997, carrying a chip from rival Cyrix Corp., Intel executives began to worry that cheap PCs could choke off its growth engine. Soon, inexpensive machines were flooding the market, and by mid-summer, Intel competitors AMD and Cyrix had grabbed 20% of the U.S. retail PC market, their highest share in half a decade.

That's when Barrett and then chief Grove decided to split Intel's Pentium line into distinct price and performance bands to target different markets. They devised the Celeron line for inexpensive PCs and the Xeon family for high-powered servers and workstations, while keeping the flagship Pentium III aimed at the middle tier of the market. Despite a cool reception at first, the Celeron has helped Intel climb back to 62% market share in sub-$1,000 PCs, up from 30% a year ago. And analysts now expect revenues from server-class chips to top $6.7 billion by 2001, up from $3.4 billion last year. The segmentation strategy has helped Intel maintain its juicy 60% gross margins by balancing high- and low-end sales.

But Barrett understood that segmentation alone wouldn't put Intel back on the growth curve investors had come to expect. Given Intel's size, growing at 20% per year required coming up with $5 billion or more in new revenues every year. And that meant Barrett had to think big. "He was looking to build Intel's next $25 billion business and knew he would have to significantly expand its charter," says David B. Yoffie, a professor at Harvard business school and a member of Intel's board.

THINKING GREEN. The CEO-to-be knew the first place to start was inside Intel. In July, 1997, he arranged for three corporate strategy gurus to conduct a seminar series for Intel's top 400 managers. In eight groups of 50, executives retreated for a week to hotels near Intel's offices in Santa Clara, Calif., Phoenix, and Portland, Ore. "We had to figure out how to grow new businesses in the shadow of the creosote bush," Barrett says.

The first day was led by Robert A. Burgelman, a Stanford Business School professor who has studied Intel for years and co-teaches a course at the school with Grove. Burgelman specializes in helping companies develop and nurture an entrepreneurial spirit. He urged seminar attendees to divide Intel's businesses into "blue" products, its old bread-and-butter chips businesses, and "green" products--everything new. The goal was to get managers to jump faster on green ideas and escape the gravitational pull of the blue status quo. Executives at Intel now routinely use the blue vs. green descriptions when discussing strategies. "There's been a freeing up of thinking," says Paul S. Otellini, the head of Intel's processor unit.

On the second day, managers learned about the threat posed by so-called disruptive technologies. Clayton M. Christensen, a Harvard B-school professor and author of The Innovator's Dilemma, explained how makers of cheap steel reinforcing bars, known as rebar, once dismissed as insignificant by industry giants, had nibbled their way into the market for higher-value steel. The PC had done the same thing to mainframe computers, he said.

"DIGITAL REBAR." The third day of the seminar focused on business ecosystems. Led by James F. Moore, president of consultancy GeoPartners Research Inc. and author of The Death of Competition, the session taught Intel execs that to move beyond the familiar PC world, they would have to construct new webs of relationships and help seed emerging business ecologies. In the first few months of the seminar series, when Intel still believed that cheap PCs would be a short-lived craze, "there was a lot of denial of the message," Moore says. Then, with AMD's market share soaring, the mood turned to despair.

The big breakthrough came in September, 1997. At a dinner with Grove and other top execs following a seminar, the famously paranoid Grove latched on to the rebar story as an analogy for cheap PCs. "If we lose the low end today, we could lose the high end tomorrow," he exclaimed. From then on, Grove referred to cheap PCs as "digital rebar," and Intel became more aggressive in promoting the Celeron, even at the risk of cannibalizing sales of pricier chips.

Intel's execs also began to accept the idea of widening the company's mission. Their vehicle of choice was an investment fund Intel launched in 1991 to dribble money into PC industry startups whose products or technologies gobbled up PC power and, in the words of Grove, created "waves of excitement" among potential PC buyers. Now renamed Intel Capital, the fund has vastly expanded its outlays, from about $300 million in 1997 to $1.2 billion last year, and has stakes in more than 350 software and Internet companies.

Intel Capital has scored some notable hits, including eToys, Red Hat, and Inktomi. Its holdings are now valued at more than $8.2 billion, and it kicked $327 million in pretax income into Intel's fourth-quarter results--half of the company's 6 cents upside earnings surprise. But Intel Capital wasn't conceived primarily as a moneymaker. The investment in eToys, for example, was seen simply as a way to boost the growth of e-tailing--and the purchase of PCs. "My goal is to expose the company to every facet of the Internet economy," says Leslie L. Vadasz, manager of the fund. "It has already lead to an opening of minds." The model worked so well that in 1999 Intel set up a separate $250 million fund to encourage software development for its new 64-bit Itanium processor and a similar $200 million fund to spur adoption of its new networking chips.

After the seminar series ended in 1998, the minds of Intel's managers were opened a bit more than the company could handle. At each session, attendees had broken into groups to dream up new business ideas. But "Barrett had gotten all these people excited with nowhere to go," says D. Craig Kinnie, director of Intel Architecture Labs, a research group in Oregon. The new CEO soon realized he needed more than just a cultural awakening: He had to change Intel's inflexible structure to allow new ideas to thrive.

WAD OF CASH. That July, Barrett came up with a solution. He and chief financial officer Andy D. Bryant threw out Intel's rulebook for funding new programs and established the New Business Group. Rather than setting tight budgets and subjecting internal startups to rigorous reviews, they were treated like venture-capital-financed companies and given a wad of cash to spend until it ran out. Under this scheme, Parker's unit has launched about 20 seed projects, each with three or four employees and a budget of several hundred thousand dollars. Seven larger projects have received $5 million to $10 million in funding. The ideas range from a scheme to equip doctors with secure IDs to encourage online medicine, to installing 3,000 information terminals on the backs of seats at Madison Square Garden for hockey fans to look up information on their favorite players. Two potential spin-offs: An Intel business called Vivonic, based in Oregon, will sell handheld computers that help users monitor their diet and fitness starting this spring. Another Portland-based venture, called PassEdge, is set to be launched this April with technology for protecting online content, such as digital music and movies, against illegal copying.

By far, Intel's biggest gamble is its nascent Web-hosting business. The idea of running Net data centers had been floated in 1996 but was shot down because top execs didn't grasp the coming rise of Web services. Two years later, in September, 1998, Barrett directed Renee James, Grove's chief technical aide, to explore the idea again. She pulled in a team of five people and spent six weeks cobbling together a plan. "We pored over reams of data, but in the end we took a flyer," James says. Intel's board quickly approved the plan in November, and by January, 1999, it was launched. Just nine months later, Intel opened its first data center in Santa Clara. With a capacity for 10,000 servers, the center is now barely occupied. So far, only about a dozen companies--including Citigroup, an Excite@Home shopping service, and several customers from Intel's Pandesic joint venture with SAP Corp.--are using the service.

Web services couldn't be further from the chip industry. But Intel execs argue that being in the business will give the company insight into e-commerce trends. Plus, it's an insurance policy in case computing moves to a pay-for-service model and PCs are eclipsed by devices connected to Web servers. Besides, Intel executives point out, the company already does $1 billion in online business every month. "We run 24 hours a day, 365 days a year," Barrett says. "We went from selling nothing online to more than $1 billion a month over our own infrastructure. We have expertise in this space."

WIRELESS WAY. Still, analysts remain ambivalent about the likely success of the program. "I just don't get it," says Manoj Nadkarni, president of consultancy, an investment research house based in Federal Way, Wash., that specializes in chip companies. Noting that the leading independent Web-hosting outfits, Exodus Communications Inc. and Verio Inc., are still losing money, Nadkarni and other analysts wonder what kinds of margins Intel can get from the business. Barrett won't comment on profits. Competitors, naturally, can't resist. "This is a mistake," says William L. Shrader, CEO of Internet service provider PSInet Inc., pointing out that Intel is going into business against its customers. "Intel looks like a joke. They'll retreat in less than two years as gracefully as possible." Adds Ellen Hancock, CEO of Web-service rival Exodus Communications: "It's a stretch for them to say they have some expertise here. We've taken years to set up our operations."

Barrett rejects such notions but admits that Intel Online Services isn't the centerpiece of his strategy. By comparison, networking and communications are "an order of magnitude" more important, he says, because the communications industry is larger than computing, yet Intel "doesn't play nearly as significant a role there." Intel, however, does stand a good chance of becoming a major supplier of wireless chip technology. The purchase of DSP Communications, combined with a joint venture with Analog Devices to develop a new digital signal processor chip, could give Intel a bigger piece of the explosive mobile-phone business. It's already the No. 1 supplier of memory chips used in cell phones. By packaging DSPC's software with Intel's energy-efficient StrongARM processor, Barrett figures he can parlay that position into selling the more profitable brains of wireless Internet-ready phones. It's a huge opportunity: Researcher International Data Corp. figures such devices could surge to 536 million units in 2003, up from 85 million in 2000. "Wireless access is the second coming of the Internet," says IDC researcher Iain Gillott.

But rivals aren't rolling over. The $7 billion communications chip sector is on fire, with projected 20% growth this year. Established players are beefing up their portfolios. Motorola, Conexant, and Lucent Technologies, for instance, all recently bought network processor startups, and IBM has a chip of its own. Competitors pooh-pooh Intel's prospects, arguing it lacks expertise in custom chips and analog circuits--both crucial for networking. "Intel doesn't understand the communications market," snaps John T. Dickson, president of Lucent Microelectronics Group, the No. 1 supplier of communications chips. What's more, says Charles Boucher of Bear, Stearns & Co., after seeing how Intel sucked the profits out of the PC makers, networking companies "are suspicious of their intentions."

Adding talent through acquisitions could help, but Intel's track record in promoting outsiders is poor: Only three CEOs from the 20 companies Intel has bought in the last three years are now among its top 92 executives. Most others have stayed, but they toil deeper in the ranks. By contrast, Cisco's top tier is populated with former heads of acquired companies. Intel acknowledges, too, that while its overall turnover remains an enviable 4% per year, it has had trouble holding on to younger, midlevel managers who are defecting to dot-coms. Says one former Intel exec: "There aren't any whiz kids there."

Barrett is undaunted. He's driving Intel at a pace it has never known before, even during the heyday of "only-the- paranoid-survive" Grove. The new chief executive is bent on leaving a legacy every bit as large as Grove's, but he's managing a far more complex enterprise facing much greater challenges. If anything, the need to prove that Intel's success wasn't just a fluke has its managers fired up. "We felt like we had succeeded before the Net," says Sean Maloney, Intel's worldwide sales and marketing manager. "Now, we have to prove it all over again." And show the world that Intel is no creosote bush.

Before it's here, it's on the Bloomberg Terminal.