Intel Corp. (INTC ) Chief Executive Craig R. Barrett gulps down an orange soda. After folding his 6-foot, 2-inch frame into a chair in a cramped conference room, it's clear he is uncomfortable with more than the tight quarters. The topic of discussion is his legacy--or what it will be. Halfway through his tenure as Intel's fourth CEO, the 62-year-old Barrett has three more years until he hits mandatory retirement age. Emblazoned on the awards, plaques, and framed articles crowding the walls, the accomplishments of Barrett's predecessors are daunting reminders of the short time he has left to satisfy his own ambitious goal of reinventing Intel.
Here, more than anywhere else, the word "legacy" smothers Barrett. The mementoes around the room mark the legendary work of Chairman Andrew S. Grove, the Midas of microprocessors who, practically by force of will and a vision of the rapid growth of the personal computer, shaped Intel into one of the world's most powerful technology companies. Grove replaced Gordon E. Moore, the technical genius who set in motion what became codified as Moore's Law: that processor performance will double every 18 months, getting cheaper along the way. Before Moore came Robert N. Noyce, an Intel founder who co-invented the integrated circuit, paving the way for countless technological advances that have formed the bedrock of the computer industry. And Barrett's legacy? "My legacy is not to be the trivia question `Who was the fourth CEO at Intel?"' he says. "I hope not to be `I don't know, who was it?"'
There's no doubt Barrett is having a hard time earning his place on the wall. After launching a bold strategy three years ago to move Intel beyond PCs and into such markets as communications, information appliances, and Internet services, the chipmaker is in its worst shape in more than 15 years. Oh sure, there isn't a tech exec on the planet who isn't having a crummy year because of a souring economy and the threat of war in the wake of the terrorist attacks. But Intel's problems run deeper than these events. For the past three years, Intel has seesawed between product shortages and product delays in its core computer-chip business. Piled on top of that have been embarrassing bugs, recalls, and overpriced processors that opened the door for rivals. By yearend, analysts expect Intel's share of the PC chip market to drop to 78%--nine percentage points below what it had when Barrett took over.
Barrett's invasion into new markets has been even more dismal. So far, some $4 billion of Intel's more than $10 billion in new investments have produced little. This year, Intel stopped making network servers and routers after some of its biggest chip customers, including Dell Computer Corp. (DELL ) and Cisco Systems Inc. (CSCO ), slapped Barrett's hands for competing against them. In February, Barrett shut down a service for broadcasting shareholder meetings and training sessions over the Web. He shuttered iCat, an e-commerce and hosting service for small and midsize businesses. And he has retreated so far in the information-appliance business that Intel now markets its Web-surfing devices only in Spain. "Certainly, Craig's vision looked a lot more attractive a year and a half, two years ago," sighs board member David B. Yoffie, a professor at Harvard Business School.
EARNINGS DROP. What went wrong? Critics say Barrett has been trying to move Intel into too many new markets, fracturing the company's focus on its core business. To execute on so many fronts, he has decentralized the organization and delegated a lot of decision-making. But getting a workable structure in place has been a challenge. Barrett has restructured the business groups at least three times in as many years, shuffling execs like cards in a deck. Even in the core microprocessor group, a startling 80% of the unit's staff were given new roles in a March shakeup. "Typically, people moving around a lot are not sure where they are going," says one longtime customer. Adds G. Carl Everett Jr., a former general manager of Intel's Desktop Products group, who left the company in 1996: "They're dabbling in everything and overwhelming nothing."
Now, Intel is bracing for its worst financial results since it fled the memory-chip business in 1985. Sure, the entire semiconductor industry is in its worst slump in a decade, suffering from overcapacity and weak demand that will cause global chip sales to tumble 34% this year, according to researcher IC Insights. But Intel will take a bigger hit, because it has failed so far to wean itself from dependence on a slowing PC business. Intel's revenue is expected to decline 52%, from $33.7 billion in 2000 to $25.5 billion this year--the chipmaker's first revenue drop since the 1985-86 tech recession. Profits are falling off a cliff, too, plummeting from $10.5 billion in 2000 to $773 million this year, estimates Merrill Lynch & Co. And it's likely to get worse. In the week following the September 11 attack, UBS Warburg analyst Don Young says retail PC sales tanked 50% year-over-year at a time when most were expecting a pickup in demand. And the news from corporate America isn't any better. Consider Frank J. Fanzilli Jr., Chief Information Officer of Credit Suisse First Boston, and what he has to say about buying new PCs under current economic conditions: "I will only be looking to replace PCs that are really crawling, and you're going to have to prove to me that they're crawling on all fours before I replace them." On Oct. 1, Merrill Lynch cut its fourth-quarter revenue estimate for Intel by 2%, to $6.15 billion.
It's now unlikely Barrett will reach his goal of 20% revenue growth in the overall business next year. For fiscal 2002, Merrill Lynch figures Intel's revenues will rise only 13%, to $28.8 billion, while net income is expected to rebound 147%, to $1.9 billion. Next year's profit growth sounds good, save the fact it will be Intel's third-worst showing since 1986. Barrett made his growth predictions before September 11, and even though he has not revised his goals, he is hinting at more job cuts. So far this year, he has cut 5,000 jobs through attrition, bringing the workforce to 80,000. In a Sept. 26 speech to employees, the CEO pointed out that Intel's head count is 20,000 employees higher than it was two years ago, even though revenues are the same. The message wasn't lost on his audience, nor on his shareholders. Before the terrorist attack, Intel's stock was hovering at $26--more than 60% off its 52-week high. Today, the stock is around $20, amid fears that shaken consumer confidence will hurt PC sales. Says a top executive at a major PC maker: "I have never seen them at a weaker moment in the history of Intel."
That's raising questions about Barrett's leadership. No one is calling for him to step down, but a chorus of former executives, analysts, customers, and partners say Grove should return as Intel's big strategic thinker. For years, Grove and Barrett had been the chip industry's dynamic duo--Grove was the visionary and Barrett was the nuts-and-bolts operations chief. "They complemented each other. In the old days, Andy was like Batman, and Craig was Robin," says one former top executive. "But everyone knows, when you want to figure out how to beat the bad guy, you don't call Robin."
Yet Grove's booming voice is heard less and less frequently at company headquarters. As chairman, he remains a force at Intel, but he says his role is mostly as an observer when it comes to decision-making. Always a cautious, methodical planner, Grove says he has been uncomfortable with Barrett's bold actions, but he defends the current CEO. "If he was equally as cautious, we would be destined to remain a niche player," Grove says. Still, Grove notes that he sought the advice of Gordon Moore during his early years as boss far more frequently than Barrett seeks him out. Barrett "marches to his own drummer," Grove says. "It's not that he doesn't take advice, he doesn't depend on it."
"NOT GUILTY." Indeed, Barrett isn't the least chastened by Intel's flagging performance or criticism of his many-pronged strategy. With a temper at times as prickly as the cactus towering over his 9-by-9 cubicle, Barrett says his strategy has not created a company without focus. "Guilty as charged, we had product screwups," he says curtly. "Not guilty as charged that we can't do more than one thing at a time." The product shortages, bugs, and recalls, he says, were "side effects" of pushing forward into many markets at a breakneck pace. Microsoft Corp. (MSFT ) Chairman William H. Gates III, who also steered his giant software company into new turf while building his core Windows business, agrees. "You can't do this without sometimes hitting a bump in the road," says Gates. "But Intel is making a lot of smart bets on the future."
Barrett says he's not backing off those bets. Three years ago, he vowed to branch out into communications, info appliances, and Internet services. His original vision not only called for making chips for networking gear, cell phones, and handheld computers but also for churning out Intel hardware--network servers, Web-surfing devices, and routers to guide data over networks. At the same time, Barrett tried to build a services business, with Intel running e-commerce operations for others or dishing up business software to corporate customers over the Net. The full scope of his vision has been far from realized. Intel has retreated from most of the Intel-branded product offerings to rely on what Intel knows best--making chips. Now, his beyond-the-PC plans translate into producing tiny slivers of silicon to go into wireless and other communications products.
Intel's problems, Barrett says, are largely the result of the economic downturn. Still, he maintains, he's turning that to his advantage by plowing gobs of money into research and manufacturing advances--$11.5 billion, a staggering 45% of revenues--at a time when rivals can ill afford such lavish spending. That's classic Intel, a ploy Grove used against cash-strapped microprocessor rivals to widen his lead in the mid-1980s. With a $10 billion cash reserve and a seasoned team, Barrett says Intel is positioning itself to come out of the downturn in better shape than rivals both old and new. The downturn, he argues, is giving Intel time to hone its next-generation products in all the markets he has targeted. "We've got the technology, we've got the strategy," he says.
That may not be enough. Now more than ever, flawless execution in Intel's core microprocessor business is essential to overcome weak results in the new businesses and the plunge in the economy. Rival Advanced Micro Devices (AMD ), however, hopes to throw a monkey wrench into that plan. The scrappy upstart, locked in a bitter battle with Intel, has made major inroads since 1999, when it introduced its Athlon processor. The chip, which was faster than Intel's Pentium III, was quickly snapped up by PC makers, especially when Intel couldn't make enough processors to meet sizzling demand during the Internet heyday. AMD, after nearly two decades, finally put a dent in Intel's market share.
NO SWEAT. This year, Intel counterpunched. It slashed prices on its Pentium 4 chips by 84%, dragging AMD into a savage price war. And in the past month, Intel has released a speedy, 2-gigahertz Pentium 4 chip that outguns AMD's top processor. AMD is expected to strike back by mid-October with faster Athlon chips. But Barrett isn't sweating it. "Our Pentium processors are coming out ahead of schedule. Our chipsets are coming out ahead of schedule. Our whole road map is accelerating," he crows.
The only problem is, the PC market isn't traveling at the same speed. This year, for the first time, PC sales are expected to decline. Even when the economy crawls back, few expect buyers will return to their habits of the past, when ever more powerful PCs were a must-have. Increasingly, customers are more dazzled by the speed of their Internet connections than by the speed of their PCs. Intel has been getting that message since January, as its chips sped to lofty new heights but sales slumped. Indeed, the average life span of a corporate PC is expected to rise to four years by 2004, from 3.3 years in 1999, according to researcher Gartner Inc. And the percentage of U.S. homes with PCs has been stalled at 58% for the past two years. "Consumers want to know what's in it for them, not just `This is a bigger number, so it's better,"' says Larry Mondry, chief operating officer of retailer CompUSA.
Barrett shouldn't expect help from his scramble into powerful, high-margin chips for servers either. Selling for up to 40 times the price of a PC chip, analysts say Intel's new Itanium server processors will take at least five years to gain traction. Customers will need to test each new chip and rewrite software to take advantage of processing in massive 64-bit chunks instead of the more common 32 bits. That's painful news for Barrett, considering Intel has staked so much on Itanium--a joint project with Hewlett-Packard Co. (HWP ) that took seven years and $2 billion yet came to market two years late, missing the tech buying boom. This year, says IDC analyst Steve Joslin, sales of Intel-based servers will decline nearly 10%, to $23 billion, from $25.3 billion a year ago, compared with a modest 3% decline in the competing Unix server market. After the terrorist attacks, Joslin revised his 2002 projections downward from 9% growth to a 1% decline. "It's not pretty out there now," he says.
And the new businesses won't be much help any time soon. In three years, Barrett has pumped more than $10 billion into 34 acquisitions to bolster efforts in new markets, betting that those deals would help such units as the Communications Group and the Wireless Communications & Computing Group grow 50% annually. Revenue did rise sharply, to $6.4 billion last year from $3.9 billion a year earlier, and now amount to nearly one-fifth of Intel's total. But that's mostly thanks to Intel's flash-memory chips, which are used in cellular phones to hold data stored in memory when a device is turned off. Even including the profitable flash-memory operation, Intel's new businesses--everything from Web hosting to processors for phones--have reaped zilch in profits, with losses doubling nearly every year since 1998.
It hasn't helped Intel's cause that Barrett has been shuffling management and reorganizing business units nearly as often as the company trots out new processors. First, Barrett formed a new wireless unit in December, 1999, combining Intel's flash-memory business and new acquisitions, such as DSP Communications Inc., a leading supplier of chipsets and software for digital cellular communications. A year later, Barrett combined the manufacturing and development groups working on its core processors into the Architecture Group. This year, Intel's networking and communications businesses were merged into a new unit, and the architecture unit was revamped for a second time in two years. "We reset the whole thing," says Paul Otellini, senior vice-president for the Architecture Group.
Barrett and Grove say the moves were necessary to clean out organizational cobwebs and to reflect the realities of a more segmented marketplace. The reorganizations were designed to make Intel more fleet-footed and to avoid duplication of efforts in various units, such as the networking and communications operations each pitching similar products to customers with no coordination. Barrett explains that in the past three years he has moved four executive vice-presidents into new positions to give them experience in sales, marketing, and operations, in preparation for choosing his eventual successor. Those on the up-and-comer list: Otellini; Sean Maloney, Communications Group; and Michael Splinter, Sales & Marketing Group. Barrett also says the reorganizations were key to pushing decision-making down into the executive ranks and making Intel nimble.
BROADBAND BLUES. Too bad Intel has been anything but nimble. Customers, analysts, and former execs say many of Intel's execution problems have been the result of putting chip experts in charge of new businesses. "They keep putting chipheads in positions they know nothing about," grumbles one former employee. Nonsense, says Barrett. He insists that Intel's managers are more than capable of tackling new challenges.
When Barrett moved into the corner office, he vowed to reinvent not only Intel's business but also its culture. He brought in consultants such as Harvard Business School professor Clayton M. Christensen, author of The Innovator's Dilemma, to encourage the troops to tear down the old ways. But Barrett has yet to purge Intel of a culture that is used to competing in markets where it is the only choice. Take Intel's much-hyped broadband effort. In 1998, Intel began pushing a new standard for a speedy, low-cost, digital subscriber line (DSL) chip without phone-company input, says Bob Merritt, director of emerging markets for chip researcher Semico Research Corp. That frustrated phone companies, which then refused to support Intel. In March, Intel quietly admitted it has no plans to ship the chip, and the future of that business is now up in the air. Barrett has worked on improving customer relations, but some say Intel's take-no-prisoners culture remains unchanged. "To be candid, Intel has been Intel," says Duane E. Zitzner, president of HP's computing-systems group. "I haven't observed any difference in the way they do business."
Barrett's answer is to turn up the heat on what he knows best: manufacturing. He has hurriedly ramped up a makeover of Intel's manufacturing plants, introducing the latest production technologies to help slash the costs of making chips and chipsets. He is pumping $7.5 billion, or 29% of expected annual revenues, into capital development. Intel is in the process of converting its plants to bring down costs of making chips as much as 50%. New, 12-inch wafers, about the size of a dinner plate, have 240% more room to produce chips than the current 8-inch ones, which are about the size of a salad plate. Copper interconnects boost performance by lowering heat conductivity, and shrinking the size of circuitry from .18 microns to .13 microns cuts the size of the same chip in half.
Indeed, Intel workers in sterile rooms wearing bunny suits soon could be a thing of the past, as virtually automated plants use sophisticated, custom-made equipment to move parts through the manufacturing process. CIBC World Markets semiconductor analyst Quinn Bolton estimates that will push the manufacturing costs of a Pentium 4 desktop chip down to just under $39, from about $60 now. Intel is betting it can use the same technology and factories to produce chips and chipsets for each market it is pursuing.
Barrett isn't finished doling out the dough for manufacturing advances. He's spending another $4 billion to research promising new technologies, such as extreme ultraviolet light to draw features for increasingly complex chips that will run at 10 gigahertz and beyond, expected in 2005. Intel also is getting ready to bring out new technology called hyperthreading that could be a significant boost to the power of its chips by allowing one processor to perform two heavy computing tasks simultaneously--as if there were two chips in the machine instead of just one. That technology will be available for servers and workstations next year and in PCs by 2003. Intel hopes to pull miles ahead of competitors, making the technical differences in its products much more pronounced, boosting sales. "I think they realize their best strategy is to out-invest and out-technology competitors," says Dell Computer President Kevin B. Rollins.
Deep pockets and manufacturing prowess, however, may not be enough to help Intel dominate the communications, wireless, and Web-site hosting markets. The chipmaker's best growth prospect outside of servers is in its wireless group, where it hopes to make inroads with its "wireless Internet on a chip" strategy. By combining its flash memory with low-power, high-performance digital and analog chips to make a single piece of silicon, Intel plans to be a one-stop shop for the guts of cellular phones and handhelds. Its edge: offering a chip that is two to five times more powerful than those of its rivals. Intel says Siemens (SI ) and Mitsubishi have signed on. The chipmaker, however, may have a tough time persuading giants Nokia (NOK ) and Ericsson (ERICY ) to abandon trusted suppliers for Intel, an unproven wireless player, says Gartner Group analyst Stanley Bruederle.
Rivals aren't waiting for Intel to gain a toehold. Industry leaders Texas Instruments Inc. (TXN ) and Motorola Inc. (MOT ) also are moving forward with next-generation offerings in the battle for the digital signal processor market, one of the key ingredients in cell phones. And they say that Intel, the interloper, has a lot to learn in a short time. "We have over 20 years of wireless and cellular knowledge," sniffs Omid Tahernia, general manager of Motorola's Wireless & Mobile Systems Div. Tahernia says there's a huge learning curve in figuring out pricing and what customers want.
A key test of whether Intel will become a formidable player in the wireless industry outside of flash-memory chips comes early next year when it begins shipping its new, low-power XScale processor. Intel hopes to convince handheld computer makers that the chip can handle large video and audio files and can be used for always-on Internet connections. And on Sept. 26, Intel unveiled a new flash-memory chip that it claims will retrieve data four times faster than standard cell-phone memory.
In the market for Internet and networking equipment, Barrett has tried to buy his way to market leadership. In 1999, he plunked down $2.2 billion--Intel's biggest acquisition ever--for Level One Communications, a maker of chips for broadband devices that speed data around the globe. Intel's early efforts read like a "How Not To" guide for business, says analyst Brad Day at Giga Information Group. Day says Intel bet on Level One, even though it already was late in developing fast ethernet networking products. The delays, coupled with Level One managers cashing out, gave competitors the opening to turn out products superior to Level One's, making the acquisition a questionable move. Indeed, Intel developed without Level One's engineers a router chip that's selling well, Day says.
WEB WOES. And the outlook for Intel's Web-hosting venture is even more bleak. Indeed, analysts say that even though Intel has poured $2 billion into this business, it should exit the market. Demand from companies wanting others to run their Web operations has fallen dramatically since the dot-com meltdown. Few Web hosters are making money. And on Sept. 26, Web-hosting leader Exodus Communications (EXDS ) filed for bankruptcy. Barrett says he is staying the course because the market will once again boom. In the past six months, Intel has forged Web-hosting and partnership deals with the BBC, Sony (SNE ), and Commerce One (CMRC ). Still, competitors say Intel's data centers are virtually empty. "We've never run into Intel in a competitive situation," says James L. Freeze, chief strategy officer at Web-hosting company Genuity.
Such tales bode ill for Barrett's legacy. Still, he has three years to reverse Intel's slide--and 36 months can be an eternity in the ever-changing computer industry. "What people will remember him for is what he did last, how the stock did last," says Lehman Brothers analyst Dan Niles. Grove agrees that the wild cards will be the economy and how well Intel executes. In the next two months, Intel's executive team will sit down and review the strategic plan that will help determine the course of Barrett's next three years. They probably won't be able to squeeze into that cramped conference room, but wherever they go to map out future plans, Intel's legacy will loom large.
By Cliff Edwards in Santa Clara, Calif., with Ira Sager in New York and bureau reports