INTEL: WAY OUT IN FRONT,BUT THE FOOTSTEPS ARE GETTING LOUDER
Four years ago, when Andrew S. Grove was promoted from president to chief executive officer of Intel Corp., he told Chairman Gordon E. Moore that he planned to retire at 55. Moore scoffed that the 55 retirement age was like the 55 mph speed limit--a number few people stick to. Moore was right. Grove will be 55 in September, but he's not going anywhere. The CEO has a new goal: the transformation of 22-year-old Intel from a personal computer chipmaker into the dominant semiconductor supplier to the entire computer industry. "I'm in a new job," Grove says. "We have to understand things that we didn't even know how to spell a year ago."
One word that he has mastered is marketing. Intel discovered the value of marketing in the late '80s, as it extended its dominance in PC chips. Marketing will be critical to Intel's efforts to extend its influence. So, on Apr. 22, when the company unveils its 486SX chip, a low-cost version of its mainframe-caliber 80486 microprocessor, it plans to back the announcement with a two-year, multimillion-dollar advertising campaign aimed at consumers. Grove figures that if Intel creates a wave of demand for machines built around the chip, computer makers won't have much choice but to use it. Indeed, IBM and clonemakers Dell Computer and AST Research plan to announce new PCs based on the 486SX within days of its introduction.
But for the long term, the campaign is designed to stimulate demand for Intel chips in every type of computer--from tiny "palmtop" personal computers to giant supercomputers. Already, NCR Corp. has bought into the concept: It plans to use Intel chips in its laptops, mainframes, and everything in between. And Grove hints that another big computer maker will soon be following NCR's lead.
If Intel succeeds, the rewards could be enormous. "With its computer engine, Intel is to the computer industry of the 1990s what IBM was to the computer industry of the '60s and '70s," says Daniel L. Klesken, an analyst with Prudential Securities Inc.
MONOPOLY. Intel is off to a flying start. It holds a 50% share of the market for microprocessors, and while the world chip business grew only 4% last year, Intel grew 25%, to $3.9 billion. Earlier this month, Intel reported that first-quarter revenues were up 27% over the year-ago quarter, to $1.1 billion. And because it holds a monopoly on its most recent chip designs (until 1986, the chips were also produced by "second sources"), profits are soaring. In the latest quarter, Intel's net leaped 37%, to $197 million. Meanwhile, the company has piled up $2.3 billion in cash and short-term investments. Its stock now trades at around 54, up from 38 1/2 in January.
In addition to its growing microchip monopoly, Intel credits its recent success to diversification efforts. Besides the microprocessors that act as the heart of most PCs, it also sells complete PCs to other companies, who put their own logos on them. The company also sells supercomputers. Intel logged $786 million in such sales last year and will likely top $1 billion in 1991. It even makes add-on circuit boards that give PCs modem or fax capabilities. Those are sold directly to consumers. Neither line is as profitable as chips, but through them, Intel gains insight into trends: Knowing what needs to go on a board this year helps it determine what should go into microprocessors next year.
Intel does have a weak spot, however, and it's a big one. The company is barely a player in the RISC (reduced instruction-set computing) computer market -- one of the computer industry's fastest-growing segments. Once aimed at scientists, RISC-based computers are cropping up as replacements for minicomputers, mainframes, and even PCs. "I can't understand why anyone would buy a PC for anything any more," says Charles S. Dvorkin, a vice-president at Teachers Insurance & Annuity Assn. in New York. He's replacing its PCs with RISC-based workstations.
ALARMING. Intel does have a RISC chip, the two-year-old i860. But it is used mostly as a graphics chip, not as the heart of RISC-based computers. Instead, chips designed by Sun Microsystems Inc. and MIPS Computer Systems Inc. are walking away with that business. Most alarming for Intel: A consortium of 21 companies said on Apr. 9 that they have agreed on a computer design based on a MIPS chip. The group's leader is Compaq Computer Corp., a top Intel customer.
Compaq CEO Joseph R. Canion insists that his company is "not evolving from the Intel platform." Still, a Compaq insider says the company decided to add a RISC line because it's worried that Intel's chips aren't keeping up. "What this consortium is saying is: 'In the future, don't look to Intel for top performance.' " If that's true, Intel's chip family could be boxed into the PC market, unable to extend into larger computers. That danger comes into sharp focus when Microsoft Corp. is factored in. As a member of Compaq's consortium, it has promised to deliver a new operating system by 1993 that is being designed to work on both Intel-based machines and those based on the MIPS chip. The effort means Corporate America will be able to use much of the software it already owns on RISC machines. Microsoft Chairman William H. Gates III denies that the plan shows a lack of faith in Intel, but he acknowledges that "it does mean Intel has to look over its shoulder more." And Intel's growing success in finished computers poses another problem: Some PC makers that rely on Intel microchips have recently made it clear that they're not pleased about competing with Intel in their own markets. While Compaq's Canion denies that it was a factor in Compaq's decision to branch out with a MIPS line, he admits he isn't happy about selling against Intel. Others, such as Safi U. Qureshy, president of AST Research, are more resigned. "We can't cut off suppliers just because they compete with us," he says.
That's especially true when there isn't much of an alternative. To wean computer makers from the 80286, which was second-sourced by Advanced Micro Devices, Intel trotted out the 80386SX chip in 1988 and refused to license the design to other chipmakers. A scaled-down version of the $300 80386, it was priced at $160 but quickly fell to $80. Intel could afford the low pricing because of the economies of scale that its monopoly delivers. That drove home a lesson: Intel could go further on high volume and low prices than with esoteric advances. "It's a business school problem, not a technology problem," says David L. House, an Intel senior vice-president.
BEST-SELLER. The 80386SX also taught Intel about marketing: Intel backed the chip with a $20 million promotion budget. The campaign was unusual for a company that sells few products directly to consumers. But it worked. Buyers began asking for PCs based on the chips, and computer companies rushed to make them. The 80386SX is now Intel's best-selling chip ever: The company will ship about 7.5 million units this year, says Prudential's Klesken. In its best year with the 80286, Intel sold just 4.5 million. "It's mind-boggling how successful they made the SX," says Gates.
Those numbers have encouraged Intel to fight the RISC chipmakers by enhancing its conventional chips rather than pursuing RISC more aggressively. In the past few years, it has doubled to four the number of design teams working on versions of what it calls the X86 family, and a fifth is in training. Intel will also spend $1 billion this year building two new plants to make X86 chips. And 75% of a planned 23% increase in research and development--to $640 million--will go to its X86 line. That will lead to six new versions of the 80486 in 1991 and to a new generation of chips, probably called the 586, in mid-1992.
First up will be two versions of the 486SX. Analysts figure the chips will start at about $400 and drop to $300 before the end of the year. By summer, says Intel's House, the company will have shipped 1 million of the two-year-old 80486. By Christmas, thanks to the addition of the 486SX, sales for the 80486 family will double, to 2 million.
Numbers like that make it tough for even the company's most serious critics to predict big problems for Intel in the short term, despite competition from RISC. In a business that progresses with lightning speed, it's too early to handicap the company's long-term plans to extend its influence. Grove has few doubts: "We are going to outperform anything anybody else will offer," he promises. "Otherwise, we are incompetent." And so far, Intel has shown very little evidence of incompetence.Richard Brandt in Santa Clara, Calif., with bureau reports