(FORTUNE Magazine) – WOULD YOU BASE your business strategy on the assumption that AT&T, IBM, Matsushita, Motorola, Philips, Sega, and Sony won't be able to keep up with you? How about gambling nearly a third of your company's revenues, $3.5 billion this year, that you will dominate a gamut of businesses in which you have no experience? Not many executives would have the stomach for such risk. But to Andy Grove it's just another gung-ho year as CEO of Intel, the company whose microprocessor chips are the silicon brains of PCs. Hyperaggressiveness has always paid off for Grove, 57, an emigre known to his peers in the chip industry as the "Mad Hungarian." By pumping megabucks into product development and factories, he has established Intel as the hardware hegemon of the PC revolution, much as Bill Gates has made Microsoft the software champ. Intel supplies the microprocessors in about 75% of all PCs sold -- including, probably, the one on your desk. Compute this: The company's gross profit margin -- revenues minus the cost of sales -- is a sublime 58%. Intel is still growing so fast it doubles in size roughly every two years. Net earnings last year were $2.3 billion on $8.8 billion in sales, making Intel the most profitable company of its size in the world. It is also, arguably, one of Wall Street's most undervalued technostocks. Intel's shares have long trailed the market; in mid-April they lost some 17% of their value in part because IBM announced a deal with Cyrix, a $125- million-a-year chip designer in Richardson, Texas, to produce imitations of Intel's products in Big Blue's state-of-the-art plants. At a recent $57.50 a share, Intel was trading at only 10.7 times earnings, vs. 20 for Standard & Poor's 500-stock index. Even the company's larger-than-life return on shareholders' equity -- 29% in the last four quarters -- hasn't persuaded investors to bet with Grove. Where many pros see a bargain -- top semiconductor analysts, including Eugene Glazer of Dean Witter Reynolds and Thomas Thornhill of Montgomery Securities, rate Intel a buy -- investors seem to see mostly peril. That's not so surprising. As almost everyone who has ever switched on a PC knows, Intel is facing greater competition than ever before. The most dramatic threat is the PowerPC, an ultrafast and surprisingly inexpensive microprocessor developed by IBM, Motorola, and Apple Computer. In March, Apple unveiled a new line of Macintoshes that incorporate the PowerPC. The machines are impressive. IBM has high hopes for its own PowerPC personal computers, which it will launch later this year. Meanwhile, Intel's market share is being nibbled away by clonemakers like Cyrix, whose microprocessors mimic Intel's and can run essential software like Microsoft Windows at a lower price. The No. 1 clonemaker, Advanced Micro Devices, won a crucial lawsuit against Intel in March, affirming AMD's right to copy Intel codes that govern how microprocessors behave. Compaq Computer, the No. 3 personal computer maker, recently announced it would start putting clone chips in some machines after years of relying exclusively on Intel. The clonemakers' growing strength has attracted even Microsoft, which since the dawn of the PC era has written almost all its operating system software to work on only Intel-style chips. Microsoft has established self-testing guidelines for chipmakers and other companies, and began licensing those that comply to market their products as "Windows-compatible." Grove claims Intel welcomes the rising competition, especially from the PowerPC: "We needed a little threat, a good target. The juice is flowing." His real response is far more fierce. He is gunning the throttle of Intel's mighty factories and labs, pushing them to a pace meant to break any competitor that tries to keep up. The attitude Grove cultivates in employees is ferocious too: It is what chief operating officer Craig Barrett proudly refers to when he says, "We're competitive paranoids." Intel people aren't crazy, though, to think many in | the industry want them to fail. Like Microsoft, the other near monopoly in PCs, Intel is widely hated and feared. Computer makers are painfully aware that while they operate on paper-thin margins, Intel is pocketing huge profits. Says Aaron Goldberg of Computer Intelligence InfoCorp, a Silicon Valley market research firm: "Intel has never been a company to be altruistic or even fair. Greed, avarice, and paranoia are its corporate culture." Critics mutter that PC producers are so dependent on its chips that Intel doesn't have customers, it has hostages. Grove is betting that PC makers and users will remain captive, especially if Intel repeatedly floods the market with ever faster, cheaper chips. To that end, he is spending the big bucks: $1.1 billion this year on R&D, $2.4 billion on capital investment. The company has laid out $750 million for a newly opened chip plant in Ireland and is paying $3 billion for two more in New Mexico and Arizona. Intel's latest microprocessor, the Pentium chip, shows dramatically how fast the company can move. The newest version packs 3.3 million transistors in a silicon square a shade more than a half-inch across. When the Pentium is switched on, information courses along circuits a mere 0.6 micron wide; 12 dozen such circuits, laid side by side, would be thinner than a human hair. The Pentium crunches data at almost twice the rate of Intel's 486DX2, today's best-seller. That's fast enough to give digital video images the resolution and fluidity TV junkies expect. What's more, the Pentium includes many components that required separate chips and wiring in earlier PCs. That means Intel adds more of the value -- and captures more of the profit -- in every computer its customers make.

The goal Grove has set for his company: to enable PC producers to double the performance of their machines at every price point every year. The first Pentiums hit the market last spring and are beginning to supplant the five- year-old 486 family of chips. (Intel expects to produce more than 30 million 486s this year, vs. at least six million Pentiums.) But if Grove has his way, Intel's next-generation chips, code-named P6 and with about twice the processing power of Pentium, will appear early next year. The P7 is due roughly two years after that. PUSHING DOWN prices faster is also part of Grove's master plan. He argues that what Intel gives up in profit margins it can more than make up in volume. "I don't give a shit what percentage margin I have," he says bluntly. "I want to increase dollar profits, and they are a product of margin times unit volume." Prices for the latest microprocessors start high -- top-of-the-line Pentiums sell for about $995 each -- but Intel cuts them by 30% per year. Pentium PCs, which cost at least $3,500 last fall, could drop to $2,000, the typical price paid by a home buyer, by the end of this year. Grove is intensely aware of what consumers and home-office users pay because he is convinced that before long they, and not corporate customers, will keep Intel's riches flowing. Corporations can no longer absorb new generations of PCs as quickly as Intel plans to turn them out. Only consumers and home-office users, in their unending quest for novelty and convenience, can provide the demand Intel needs. So Grove is betting the ranch that Intel can take over your home even more completely than it has already conquered your office. He aims to transform the PC powered by Intel chips into an all-purpose consumer device that will ultimately subsume your TV, telephone, VCR, answering machine, videogame console, and set-top cable TV box -- and serve as your family vehicle for tooling down the info highway. That ambition could lead to collisions with the likes of Sega and Sony, which make money on single-purpose gadgets like game players and TVs. But for Intel the greater danger is in going too slowly. Unless it achieves the mass- market volumes a consumer product provides, the company eventually won't be able to cover its escalating capital costs and keep the profits coming. There's even a real risk that if Intel slowed its spending, it could begin to lose its grip on plain old PCs as clonemakers gain ground. Gonzo as Grove's strategy may seem, technology experts pronounce it sound. Thornhill of Montgomery Securities thinks that with its microprocessor near monopoly, Intel is able to exploit the market in ways other companies only dream of: "Intel is masterful at working all the strategic levers -- production capacity, timing, technology, and pricing." Says Michael Slater, publisher of the Microprocessor Report in Sebastopol, California: "The more I look at Intel, the stronger it seems." Slater is the one expert to whom executives from Intel and the PowerPC consortium defer when asked to reconcile their wildly conflicting claims about the merits of their chips. To get his vision across to the company's 29,500 employees, Grove has boiled down his goals to two lines. He walks around Intel's Santa Clara, California, headquarters handing out fortune cookies with the strategy printed on slips inside. Somewhat cryptically, the message reads in its entirety: 1: Job 1 2: Make the PC "IT" Line 1, Grove explains, means to strengthen Intel's No. 1 position in the microprocessor market and establish Pentium as the best-selling microprocessor faster than any in history. Line 2 is Grovespeak for turning the PC into the cornerstone of 21st-century information technology and using it to transform the consumer-electronics markets. IN ITS LUST for the consumer dollar, Intel is spending mightily to make itself a household name. In 1987, the year Grove became CEO, the company's advertising and promotional budget was minuscule and was directed solely at computer makers and software developers. Two years later Intel spent $5 million on its first advertising aimed at PC buyers. The ad budget has ballooned to more than $100 million this year, about half of which pays for TV commercials in the U.S. and Europe. Intel also subsidizes the campaigns of PC manufacturers that agree to emblazon their products and ads with the bright blue and white logo intel inside. According to Intel studies, consumer recognition of the logo now slightly exceeds that of the NutraSweet swirl. Grove says there are other signs that investing in consumer ads is paying off. Buying behavior has started to change: People who purchase PCs are much more eager than in the past to go for machines that incorporate the hottest new chips. In a 1991 survey, Intel found that fewer than 5% of potential customers wanted a computer with the 2 1/2-year-old 486. But late last year, just six months after Pentium's debut, 23% of those surveyed wanted the hot new box. Intel is cooking up fascinating ways to whet consumers' appetites for high- powered PCs. It conducts this research hundreds of miles from Silicon Valley at a Hillsboro, Oregon, technology hatching ground known as the Intel Architecture Lab (IAL). The three-year-old lab, which has an annual budget of $100 million, has already outgrown a collection of low-slung buildings surrounded by rye and wheat fields; Intel is building a five-story addition nearby. IAL's projects are hardly what you'd expect of a chipmaker: of 700 employees, 600 are developing software, mostly aimed at adding consumer- oriented features to PCs. Among the most promising:

-- PHONE PCs. In collaboration with Microsoft, IAL has defined standards that allow a PC equipped with a speaker and microphone to double as a programmable telephone. Now Intel is developing software that will endow Pentium computers with the power to handle telephone functions. Soon you'll be able to connect a phone line to the back of your PC and have it set up your conference calls.

-- TV MODEMS. New circuitry from IAL allows a PC to connect to a cable TV line and serve as a superintelligent combination TV and cable box. Intel says several cable systems will begin testing the modems later this year.

-- ON-LINE NAVIGATION. Engineers and product designers at IAL are working on software to help PC users cruise the notoriously labyrinthine Internet with ease. They're also developing ways for PCs to automatically filter and organize information that will pour into them via high-capacity phone lines or cable networks. Choosing from on-screen menus, you'll be able to call up data services or music, do your shopping, and order movies.

-- DESKTOP TV. Working with Turner Broadcasting's CNN, Intel's lab has cooked up a way to distribute newscasts and other videos to users of PCs in an office network. Unlike previous systems, IAL's ties up only a tiny fraction of the network's capacity -- so the TV shows won't interfere, technically at least, with other work. Field trials will begin in May.

-- WORK SHARING. A key IAL effort has already led to a commercial product: software called ProShare, which allows two PC users connected by a telephone line to work together onscreen. With ProShare, either PC can "host" the other, so the users see exactly the same data and share the same program. If the phone line linking the machines is a high-speed ISDN (Integrated Services Digital Network) line, ProShare also lets the users conduct a video conference as they work, with the help of miniature cameras mounted on their PCs. Price for the software: $200 for the version that handles documents and spreadsheets, up to $2,000 for the videoconferencing version. Intel sells the cameras too. While working on software for tomorrow, Intel is pushing Pentium hard today. Ken Lowe of Dataquest in Silicon Valley estimates that Intel will derive more than one-third of its revenues from consumers and home-office users within two years. By then Intel will have fought two or more bloody rounds of upgrades with its PowerPC and clonemaker foes. In Grove's main R&D labs -- complexes in Santa Clara and in Oregon near IAL that together employ 1,000 people / -- engineers are developing microprocessors at more than double the rate of the past. Intel used to work on one chip at a time, starting a new one only when the last was out the door. But in 1990 it began to overlap development for multiple generations; now it's designing no fewer than five. Despite the recent hubbub about the clonemaker threat, security analyst Daniel Klesken of Robertson Stephens in San Francisco expects Intel to leave its rivals in the dust. By 1996, he projects, clonemakers will have captured some 25% of the market for Pentium-class chips -- but by then Intel will be minting money with the P6 generation, for which it will have no direct rival. Says Klesken: "Intel will own the leading edge and share the trailing edge with half a dozen competitors." The Cyrix M1 chip, an elegant Pentium- compatible design, is at least nine months away from production; by then Grove intends to introduce Intel's next-generation P6 and make Pentium yesterday's news. Intel has no such technical advantage when it comes to the PowerPC, however. Nothing makes Intel executives more competitively paranoid than the new IBM/ Motorola chips. The first PowerPCs, the 601 and the 603, are smaller than Intel's best, cheaper to manufacture, and comparable in power; a newer version, the 604, is faster than any Pentium on the market and should be similarly priced. Within the next year, predicts microprocessor guru Slater, the PowerPC family will develop a performance advantage over Pentium of more than 50%. That's because the PowerPC incorporates an architecture called RISC (Reduced Instruction Set Computing). RISC microprocessors need fewer transistors to carry out the same processing chores as the CISC (Complex Instruction Set Computing) chips made by Intel. Intel has faced challenges from RISC chips before, but none have had such potent backing as the PowerPC. ALL THE SAME, Slater believes, Grove shouldn't be losing sleep. Intel remains unique in its ability to build the fastest processors for Microsoft Windows, the software that people most want to use, and the thousands of applications programs that work with it. On a non-Intel-style chip like the PowerPC, Windows generally works only under a procedure called emulation, which drastically slows performance. Says Slater: "There really is no question that PowerPC is a superior architecture. What protects Intel is that the software advantage is quite large." In a move that will only strengthen Intel's position, Microsoft is set later + this year to unveil a new version of Windows, code-named Chicago. It will be a hot product, say preliminary reports in the computer press, because Microsoft has finally succeeded in its eight-year quest to make Windows as appealing and easy to use as Apple's Macintosh. The good news for Intel: Like today's versions of Windows, Chicago will run efficiently only on Intel-style chips. Carl Stork, the Microsoft executive in charge of working with hardware companies, says Microsoft has no intention of rewriting Chicago so that it will also work on the PowerPC. THE RIVAL Intel should fear most is IBM. The world's leading maker of personal computers and No. 1 customer for Intel chips, Big Blue has a powerful incentive to horn in on Intel's profits. The agreement with Cyrix should enable the IBM microelectronics division to become a major supplier of Intel- style chips. In the eyes of many experts, the division, which until recently sold solely to other parts of IBM, is the only supplier with the skill and capacity to churn out large quantities of chips that match Intel's best. Meanwhile, Big Blue is the big force behind the PowerPC. IBM pioneered RISC architecture in the 1970s; the PowerPC is largely a product of its R&D labs. Executives have led the world to expect that their soon-to-be-launched line of PowerPC personal computers will be the first major win of the era of Lou Gerstner, who became CEO just over a year ago. Says Jim Cannavino, corporate senior vice president for strategy: "We're taking a systematic approach to make this the next generation of computers in the marketplace." IBM has its own version of the "PC is it" vision. What will differentiate its new machines is the ability to perform what IBM calls "natural computing." Cannavino defines that as a response to users' demands for no-hassle operations: "Get this computer out of my face so it's actually paying attention to me!" For example, says Phil Hester, an IBM engineer who helped master-mind the PowerPC design, look for IBM's PowerPersonal PCs within the next year to be able to handle continuous speech recognition. Translation: the computer will understand a wide range of commands spoken in a normal tone of voice without pauses between words. The systems should also allow users to easily juggle video, still images, music, and other audio. If IBM delivers on such promises, it could attract millions of new customers to the PowerPC. But there is no simple way to run the Windows software people are most familiar with on the PowerPC. Apple too would love to steal customers from the makers of Intel-based PCs but represents much less of a threat. Its Macintoshes, which require software that doesn't work on Intel machines, currently account for about 12% of the personal computer market. Could Mac's share grow dramatically? Probably not: Apple has had superior products for years, with features Microsoft and Intel still struggle to copy. Yet the low prices and wealth of choices in the much larger universe of software written for Intel chips have consistently appealed more to customers than what Apple offers. The Cupertino company could gain market share if CEO Michael Spindler makes good on a promise last January to license other computer makers to produce PowerPC machines that are Macintosh-compatible. That would give users more opportunity to opt out of the Intel/Microsoft world. But even then it would take years to significantly dent Intel's position. Eventually the distinctions between the Intel chips and the PowerPC will blur. Michael Slater believes that the RISC approach is so superior that even Intel will embrace it. Slater expects Intel to make the move by building a hybrid chip, probably the one code-named P7. It will pack both CISC and RISC circuitry so that it will accommodate software written for earlier Intel chips while providing RISC speed for new applications. The PowerPC partners are researching a similar path for at least one of their future chips, which might include circuitry to run Windows. By 1997 there could be a showdown in the market between two chips with similar capabilities. While Intel races IBM and Motorola to lead the world to ever-greater heights of computing power, most clonemakers are gambling that demand for 486-style chips will stay strong. AMD recently contracted to build two million 486s a year in an underused Digital Equipment Corp. plant in Scotland starting next fall. This year AMD will be able to build only about four million 486s, less than one-eighth Intel's production. So far, the clonemakers have succeeded best in niches. For example, Texas Instruments sells an energy-efficient 486 imitation chip that is especially suited to laptops like the ultra-lightweight OmniBook sold by Hewlett-Packard.

AS FAR AS GROVE is concerned, the threats facing Intel, most notably the PowerPC, only help it gird for what he calls the "megabattle," when he'll face off against the world's biggest electronics manufacturers. "What I'm after is televisions and telephones and every single-purpose appliance," he says. "The best way for us to go for the 250-million-unit market is to move video telephony and conferencing and entertainment and information access onto the PC and render those other things less and less relevant. Would we do it as fast without PowerPC? Truthfully, no. We are making gutsier moves investment- wise, pricing-wise, every way, because we've got a competitive threat. The net result is we'll get to advance to the next level of competition." Computer users should be grateful that such a competent and resourceful bunch as IBM, Motorola, Apple, and AMD went after Intel's business. The world will get much better PCs as a result. And the chances are that they'll have Intel inside.

CHART: NOT AVAILABLE CREDIT: FORTUNE CHART/SOURCES: COMPANY REPORTS, WORLDSCOPE CAPTION: INTEL Santa Clara, California