) chief executive and half a dozen of his top execs hit the snow for a cross-country skiing jaunt. Kicking off a January, 2003, weekend retreat, the outing soon turned into a grim slog. With the sun setting and a winter chill deepening, the guests wondered aloud whether it was time to head back to the ranch. But Barrett showed no signs of ending the adventure. For three hours, he plowed ahead at a pace that left even the fittest of the team bone-weary. For Barrett's lieutenants, the message was clear: Turning back was not an option.
Just 15 months before he hands the CEO job over to President Paul S. Otellini and becomes chairman, the 64-year-old Barrett is pushing just as hard to secure his legacy. After a rocky start to his tenure in which Intel invested billions in new businesses that largely failed, Barrett has been racing to transform the chipmaker. He has since shaken up Intel's insular, engineering-driven culture. He has ignored the scorn of critics and plowed $28 billion into cutting-edge plants and new technologies during the longest downturn in the chip industry's history. Today, Intel's manufacturing, always a cut above, is unrivaled throughout the semiconductor industry. "Everybody thought Barrett was crazy," says money manager Snehal Vashi of Henssler Equity Fund (HEQFX
), which counts Intel shares among its $1 billion in assets. "Rather than pulling back, he invested more, and that is bearing fruit."
Now Barrett is planning a last hurrah that may have some questioning his sanity all over again. At a time when some execs would be eyeing their pensions, Intel's chief is launching the most ambitious move beyond computers in the company's 35-year history. Forget Intel Inside. Think Intel Everywhere. Under Barrett's plan, Intel's powerful lineup of chips would be the guts of nearly every type of digital device on the planet. Cell phones. Flat-panel TVs. Portable video players. Wireless home networking. Even medical diagnostic gear. All told, the company is targeting 10 new product areas for its chips, primarily in the consumer-electronics and communications markets.
Barrett is convinced now is the time to strike deep into new territory. Communications, content, and consumer electronics, after years of false starts, are rapidly evolving from the old analog world to one of standardized digital products, the realm of Intel's power. Gung-ho consumers, surging broadband adoption, and rapid declines in the costs of hard drives, chips, and other key technologies are driving much of the world to digital content. Once photos, music, and video take digital form, they become the bits that Intel's chips can process, store, and zap across the Web. "Everything in the world is going digital," Barrett says. "Communications is going digital. Entertainment is going digital. We are able to bring our expertise into differ- ent areas where we really had no unique capability before."
Translation: Barrett fully intends to upend the status quo in communications and consumer-electronics markets. Think of Intel as a silicon arms dealer. By hawking cutting-edge digital technologies, the chipmaker will offer weapons that upstart companies can use to knock existing players for a loop, while forcing entrenched companies to consider buying the same weapons. This fall, for instance, Intel plans to roll out a chip based on a technology called WiMax that could be used to deliver high-speed wireless Internet access throughout a small city for about $100,000, one-tenth the cost of rolling out fiber-optic lines today. Either cable and phone companies buy into Barrett's vision, or their near-monopoly on broadband could be cracked by upstarts using WiMax. That's just the tactic that led to the dizzying popularity of Wi-Fi, a similar wireless technology, after Intel got behind it last year.
If Intel succeeds this time, the payoff could be huge. The consumer-electronics, wireless-handheld, and communications-equipment markets that Intel is targeting already use $77 billion worth of semiconductors, and Intel has less than 6% of that total. If the company can boost its share to 10% in five years, as some analysts estimate, its revenues from the new markets would soar from $4.4 billion now to about $10 billion. Add that on top of the 11% growth expected in Intel's core computer-chip business, and Intel's revenues would rise by the 15% annual average that Barrett is targeting. "They're one of the few companies in the world that can dream of doing something that massive," says analyst David Lytel of the investment research firm Precursor Group.
Sound risky? A bet-the-company gambit? At first blush, maybe. After all, Intel is investing about $2 billion in everything from chip design to marketing to get into these new markets. But look closely, and you'll see how shrewd Barrett's strategy is in limiting the risk. As Intel pushes into new markets, it's doing as the Romans did and making sure that all roads lead back to the heart of its empire: the computer-chip business. If the new markets take off, they'll pump up demand for PCs and servers -- bringing in new revenues for Intel even if it doesn't win big in those new markets. Take portable media players, which store TV shows and movies on a hard drive. If demand for them booms, Intel benefits because consumers will want beefier PCs to handle all the digital video they'll be loading onto their players. If Intel sells the chips in media players, too, Barrett grabs a share of the trade in today's Carthage and Rome. "With Craig's strategy, you get a double hit," says Charles E. Young, an Intel board member since 1974 and University of California at Los Angeles Chancellor Emeritus.
Even if one or two initiatives fizzle, Intel's world won't come crashing down. The company's core PC and server business should continue to produce solid results in the years ahead. A strong showing in the computer businesses is the primary reason Intel's profits are expected to jump 46% this year, to $8.2 billion, as revenues rise 15%, to $34.7 billion, according to investment bank Lehman Brothers (LEH
). Still, those results will be boosted by an expected recovery in tech spending. Without help from new markets, Intel risks being chained to a slower-growth PC business, especially in light of competition from a rejuvenated Advanced Micro Devices Inc. (AMD
There is reason to be skeptical about Intel's chances in several of the new markets. The company's past attempts to expand beyond the computer market have, with few exceptions, failed. In 2001, Intel backed out of making cameras and other consumer-electronics gear after key customers, including Dell Inc. (DELL
) and Hewlett-Packard Co. (HPQ
), complained that it was competing against them. And in 2002, the chipmaker took a $100 million charge when it pulled the plug on a foray into Web hosting.
Now, several of Barrett's bets are on markets that depend on unproven technologies. Portable media players, for instance, may never catch fire beyond a handful of eggheads and movie buffs. "Where they have the world's best understanding of the PC market, they may not have the best understanding of [consumer electronics]," says Gerard Kleisterlee, CEO of Royal Philips Electronics, the Dutch giant that makes both semiconductors and consumer electronics.
Even in promising areas where Intel has made headway, there have been recent gaffes. Last year, the chipmaker tried to push through a nearly 40% price increase for flash memory, the chips that store data after devices like phones or PDAs are turned off. But rivals didn't follow suit, and analysts estimate that Intel forfeited nearly $1 billion in sales as Nokia and other cell-phone makers bought memory from rivals Samsung and AMD. Intel then tried to take on Texas Instruments Inc. (TXN
) in the market for digital signal processors, one of the key ingredients in cell phones. But cell-phone manufacturers stuck with TI, since they didn't think Intel's product was significantly better. Intel admits it misjudged the market, and rivals say Intel has a steep learning curve outside PCs. "What's the difference between us and Intel?" asks TI's incoming CEO, Richard K. Templeton. "We have customers."
Maybe so, but Barrett knows a thing or two about fortitude. After sprinting up five flights of stairs at Intel's Santa Clara (Calif.) headquarters with more vigor than men half his age, he sits in an executive conference room and explains why Intel will succeed where it has failed time and again. Working with Otellini, he says he has examined the threads of failure from the past, and the two have woven together a strategy that avoids the old mistakes. He claims the chipmaker has become more market-savvy and better equipped to pick its fights in new markets. It has forged closer ties with customers, working to create products that fit their needs instead of designing products no one has asked for. And with $13 billion in its coffers, one of the largest stashes in the tech industry, Intel can pour cash into sales and marketing, he says, persuading potential partners to take its plans seriously. "You start to generate a lot of enthusiasm to get the whole ecosystem lined up," Barrett says.
Building an ecosystem is not something Intel leaves to chance. The company's venture-capital arm, Intel Capital, increasingly is being used to seed outfits that help create a cottage industry around a new Intel product. Take Wi-Fi. Many months before the chipmaker launched its Centrino product, which combines the low-power Pentium M processor and a Wi-Fi chipset, Intel Capital dedicated $150 million to backing companies that supported the wireless technology. Among the beneficiaries was Cometa Networks Inc., which created "hot spots," or places where people could get speedy Wi-Fi links to the Net. Intel followed up the investments with a massive $300 million global marketing campaign that set the industry abuzz. Now Intel Capital is using the same approach in its 10 target markets. On Jan. 7, the venture unit, which holds stakes worth $1.3 billion in 344 companies, created a $200 million fund to accelerate innovations in the digital home.
What may help that seed money take root is that several customers say Intel is becoming more cooperative. In the past, the chipmaker tried to establish standards for new technologies on its own. But recently, it joined several standards-setting groups, including one for home-electronics gear and another for wireless technology. In one case, Intel even agreed to support a mobile-phone technology developed by Nokia Corp. "It's indicative of a difference in how they're approaching the market," says Anssi Vanjoki, manager of Nokia's multimedia group. Says Samsung Electronics Co. Senior Vice-President Chang Il Hyung: "Intel has certainly softened its stance and become a cooperative partner in new markets."
More than anything, it's Intel's manufacturing that makes it a fearsome force this time. In late 2002, Barrett debated with his board because he wanted to keep investing heavily in new plants while hanging on to all of Intel's older plants. If the long downturn persisted, some board members feared, that could burden Intel with too much capacity. During a series of board meetings that fall, Barrett convinced the doubters.
Now Intel looks ready to wage its own campaign of shock and awe. By 2005, five new factories will be able to churn out chips using 12-inch wafers the size of a dinner plate, printed with 90-nanometer circuit lines just 0.1% the width of a human hair. Each plant will slash chip costs in half by producing 2 1/2 times more product than older-generation fabrication plants. That gives Intel the ability to produce 1.25 million processors a day -- a staggering 375 million a year. Because of the expense of these fabs, rivals TI, AMD, and IBM (IBM
) will have only one plant each with such advanced equipment in that time period. TI concedes that its capacity to produce the latest-technology chips will be limited to about 250,000 chips a day. "No one can go toe to toe with Intel right now," says analyst Joseph Osha of Merrill Lynch & Co. (MER
), which has done investment banking for Intel in the past.
Although Intel is taking a financial risk by investing so much in capacity, it's modest. The company's leading-edge factories will be kept running full tilt, since they're the most efficient and most capable of producing the cutting-edge chips for PCs, servers, and Intel's other established markets. Intel's seven older plants are the ones that will be used to supply the new markets, as well as commodity products like memory chips. In the worst-case scenario, with weak demand in most of the new markets, Intel may have to shut two or three of these older factories. Since the plants and equipment were paid off when they made PC chips, there would be no write-offs. Instead, Intel would sell off the gear and shutter the facility, a process the company says costs no more than $5 million per plant.
Most challenging among the new markets will be communications, where Intel's past arrogance alienated potential partners. In December, Barrett removed Ronald J. Smith as head of Intel's Wireless Communications Business and folded the unit into the Communications Group, headed by Executive Vice-President Sean M. Maloney. Barrett's third in command now has the monumental task of turning around the combined operations, which together lost $858 million last year on $4 billion in revenues.
After the success of Wi-Fi, Maloney's next great hope is for WiMax. The wireless technology is similar to Wi-Fi in that it provides high-speed Net access to computer users anywhere within the range of an antenna. The difference is that while Wi-Fi's range is 200 feet, WiMax's range extends to some 30 miles. If the technology works as billed, a company could put a WiMax node on an existing cellular tower and make service available throughout metro St. Louis or St. Petersburg for as little as $100,000. "With its much lower costs, it catches the vision of all the operators," says Zvi Slonimsky, CEO of Alvarion Ltd. (ALVR
), a WiMax equipment maker.
Trouble is, vision doesn't pay the bills. The major telecom players, including Verizon Communications Inc. (VZ
), have been building broadband networks for more than a decade using their existing copper wires and fiber-optic lines. To serve customers who want fast wireless connections to the Internet, most of the telecom giants have picked out Third Generation, or 3G, technologies that compete with WiMax. Verizon's chief technology officer, Mark A. Wegleitner, says the company tested WiMax in Virginia and found it "satisfactory." But, he says, the company has no plans to use the technology, since they've already started writing checks for 3G gear.
Other companies may be more open to WiMax. Qwest Communications International Inc. (Q
) doesn't have its own wireless network and hasn't invested in third-generation gear. CEO Richard C. Notebaert says he's looking carefully at WiMax, particularly because providing broadband over copper phone lines is expensive in much of the Denver carrier's sparsely populated territory. Startups in the U.S. and abroad may gravitate to the technology because, for the cost of one or two towers, they could compete with phone and cable companies in selling broadband. AT&T (T
), Sprint (FON
), and BellSouth (BLS
) also joined the WiMax Forum recently to keep tabs on the technology, although none has committed to deploying it.
Barrett may find it easier to shake up the $100 billion consumer-electronics industry. Traditionally, heavyweights such as Sony Corp. (SNE
) and Philips have designed their own parts, then constantly tinkered with the technology. Besides costing loads of dough, that process can give customers headaches if they purchase gear made by different companies and want everything to work together.
Now, Intel plans to change those old ways. The company is pushing for more standardization and has persuaded companies such as Samsung to use similar parts and software in their products. That approach makes products easier to use -- and gives the chipmaker the opportunity to sell more processors, sound, and graphics chips for digital TVs, cameras, and portable video players. Intel is counting on the faster product cycles and distribution capabilities of traditional PC partners such as HP, Gateway, and Dell, as well as a growing collection of Asian contract manufacturers, to give it a leg up. "Intel is a key partner of ours, and they're going to be a key player in the consumer space because so many of the consumer devices are based on Intel chips," says HP CEO Carleton S. Fiorina.
Intel's prospects look bright. Take Barrett's bet on a high-definition display technology called liquid crystal on silicon, or LCOS. Intel is committing nearly $500 million to producing the chips, whose mirror-like surfaces reflect light to produce a digital image. Execs promise Intel's manufacturing will help slash large-screen costs in half by the end of this year, to less than $2,000 for a 60-inch projection-TV set. If Asian TV makers or HP make sets using the technology, Intel could muscle in on TI and even Sony. "It was a peaceful marketplace. Now people from the outside are coming in like hunting tribes," grumbles Sony Electronics President Hideki Komiyama.
Even as Intel pushes into new markets, it will need to defend its flank. Last April, rival AMD launched its new Hammer line of microprocessors, hoping to break Intel's lock on the market for server chips. Most of today's chips process data in chunks of 32 bits at a time, but the Hammer chip for servers, called Opteron, also can process data at double the rate with no performance trade-offs. Opteron quickly gained a foothold in the server market, and Intel's rival is hungry for more. "We are going to reinvent our position in the market," says AMD CEO Hector de Jesus Ruiz.
To counter the threat, Intel on Feb. 18 revealed it is adding 64-bit capabilities to its popular Pentium and Xeon chips. Pressure from key server makers HP and Dell forced Intel to move up the 64-bit introduction by at least six months, according to two sources at Intel and HP. Intel's move could marginalize its other 64-bit chip, Itanium 2, which is targeted at high-powered server systems. Intel says it will activate the 64-bit capabilities in its new Pentium 4 chip as soon as PC software that requires the feature is available. "You can be fairly confident that when there is [64-bit] software, we will be there," says Otellini.
Soon enough, it will be Otellini who will be making the strategic decisions at Intel. A 30-year veteran who has focused on sales, marketing, and the PC business, Otellini doesn't have the deep roots in manufacturing that Barrett has. Still, he vows that he'll continue to bet just as big on manufacturing as the company's other leaders. Even a junior employee, he says, understands the value of Intel's manufacturing leadership. "And there's 45,000 other manufacturing people to help nudge me if I don't remember that," he says, laughing.
There's also Barrett. In 2005, he is expected to take over as Intel's chairman from co-founder Andrew S. Grove, who will become chairman emeritus. For many more years, Barrett will be around to argue that, like his guests on the snowy turf, Intel must slog through the semiconductor boom-and-bust cycles and focus every ounce of energy on growing new businesses. The alternative is being left out in the cold. By Cliff Edwards
With Moon Ihlwan in Seoul and Andy Reinhardt in Paris