(FORTUNE Magazine) – After the close of trading on January 14, the world's leading manufacturer of computer chips stunned no one by announcing record 1996 earnings of $5.2 billion on sales of $20.8 billion. Shareholders have grown accustomed to that kind of performance from Andy Grove's Intel. Since Grove took over as CEO in January 1987, Intel's average annual return to investors has been an astounding 44%.

More revealing than the earnings report was an executive-suite shuffle announced a day earlier. On May 21, Grove will become chairman, replacing Gordon Moore, who founded the company with Robert Noyce, the co-inventor of the integrated circuit. The shift is largely symbolic: Moore will continue to work three days a week, and Grove will continue to run the place, as he has for years. But the announcement did make one thing perfectly clear: Andy Grove now operates in no man's shadow.

To understand the significance of the move, you must know that in the storied past of Intel, one idea looms largest: Moore's prescient prediction in 1975 that the power of a computer chip would double every 18 months. What CEO Grove has done for the past ten years is to deliver brilliantly on the promise of that "law." Intel has consistently been the company to supply PC manufacturers with the hot chips they need to power each generation of new, exciting PCs. Its competitors--among those that have survived are Advanced Micro Devices (AMD) and Cyrix--are little more than also-rans. Intel microprocessors power more than 80% of all PCs in the world.

But Moore's law is not a law of physics. It is, says Albert Yu, Intel's general manager for microprocessor products, "a law of technology and business." So far, Moore's law has been sustained by a regular cycle: Computer makers and software companies (especially Microsoft) develop new features and programs that require more power. Intel, meanwhile, creates brawny new chips to meet those new demands. Selling chips at prices high enough to earn gross margins now hovering around 60%, Intel uses its profits to build new chip factories, which presently cost some $2 billion a pop--thus readying itself for the next turn of the cycle. Predictable enough in recent years, yes; immutable, no. Says Yu: "This thing could stop. If people don't buy a chip with more functions, there's no money to develop the next round."

The famously worried Grove--author of the current business best-seller, Only the Paranoid Survive--came to appreciate this fully only a few years ago. He decided that basing the business on the premise that others would create demand for microprocessors was suicidal. What if Microsoft, say, shifted its R&D from developing next-generation software to making its current products work more efficiently on today's PCs? Fewer people would want new machines powered by Intel's hot, new, expensive chips.

So Grove embarked on a radical approach to stay ahead of competitors and keep Intel growing. In recent years he has molded the giant chipmaker into much more than a supplier of parts. It now aims to be the visionary leader of the entire computer industry. From here on, Grove declares, Intel will create the demand. He explains: "If we don't make computers more useful, there won't be demand for the chips we'll be making in a few years. So we have to create users and uses for our microprocessors. We get the market growth we earn--by our development efforts, our investments, and our proselytization. That's absolutely in our psyche now."

Intel's transformation is a personal triumph for Grove, 60. Even though he was one of the handful of employees who got the company up and running in 1968, Grove has never been considered a founder--a status reserved for Moore and Noyce. They have been celebrated as legends and visionaries, aided in their success by Andy Grove, the efficient manager. But now he must be considered their equal. Says Regis McKenna, the Silicon Valley marketing pioneer who worked closely with Intel's management in the early days: "Intel in the past was fairly conservative. But today it's much more willing to take leading-edge positions, because Andy realized Intel has the power to create the market. He may be the best manager in the world. He has achieved as a manager the stature held by Noyce and Moore in the past as inventors."

For the past five years Grove has pursued a series of projects that seem to have little to do with making chips. He's been spending time in Hollywood, chatting with the kind of moguls that hobnob at Herb Allen's Sun Valley, Idaho, summer get-together; quietly investing in a range of small companies that do things as esoteric as construct 3-D interactive "worlds" on the Internet; even trying to help Starbucks set up a videoconferencing network for its customers. Grove's peregrinations have sometimes seemed puzzling, but now the reason for them is clear. To ensure Intel's success over the coming five years, he has been maneuvering to make the PC the central appliance in our lives. In Grove's vision, we will use PCs to watch TV, to play complex games on the Internet, to store and edit family photos, to manage the appliances in our homes, and to stay in regular video contact with family, friends, and co-workers. If Grove's vision comes to pass, Intel will thrive. If it doesn't, Intel's strategy falls apart.

"Intel is on a treadmill of new-product introductions fed by increasing demand for microprocessors," says Scott Randall, a security analyst at SoundView Financial in Stamford, Connecticut. "The day that treadmill slows down is the day their business plan has to be rethought." That's why Grove has boosted the budget for projects that contribute to market development but have nothing directly to do with microprocessors. Such spending has gone from zero in 1990 to more than $500 million in 1996. Intel is the only company in the computer hardware business that can afford that kind of money: Its earnings exceed the aggregate profits of the top ten PC manufacturers combined.

Intel does exactly one thing that no competitor can match: build state-of-the-art microprocessors in quantities of tens of millions. Intel continually improves its chip designs; more to the point, it is the only manufacturer that can afford to keep building the gargantuan plants required to make them. In 1996, Intel spent $5 billion on capital projects and R&D; every nine months or so it puts up a new chip plant, or "fab," each a $2 billion bet on the future. Explains Craig Barrett, who was named president on January 13: "We build factories two years in advance of needing them, before we have the products to run in them, and before we know the industry's going to grow." Says Grove: "Our fabs are fields of dreams. We build them and hope people will come."

Not that Grove lacks a road map of the technological future. At the Comdex computer trade show in Las Vegas last November, he stood before 7,000 attendees and confidently described the Intel chip of 2011. Today's Pentium Pro microprocessor contains 5.5 million transistors; Intel chips then will pack a cool billion. Today's top clock speed of 200 megahertz will soar to ten gigahertz (or 10,000 megahertz). Last year Intel sold about 60 million processors; by 2011 it hopes to be selling many hundreds of millions annually. Says Michael Slater, publisher of Microprocessor Report: "I see no clear technology threats. The biggest long-term threat to Intel is that the market growth slows."

That is the threat that worries Grove. When Intel introduces a new microprocessor (the Pentium Pro is the latest), chips from the first wave of production sell for as much as $1,000 each. They power top-of-the-line PCs and servers, which tend to be purchased by early adopters, technophiles who just can't wait. As Intel increases production, it lowers the price, stopping somewhere just north of $200 when the chip is the hot mass-market processor. That figure is low enough to ensure unit sales in the multimillions but high enough to secure huge margins.

Even at that price, Intel chips are expensive enough that PC manufacturers must charge at least $1,500 per machine to earn a profit. Says Slater: "Intel's business model is dependent on selling processors for in excess of $200, so they are also dependent on $1,500 PCs being predominant." Consumers will spend that kind of money only if they're persuaded that a new PC will help them do very cool things.

That's one reason Intel needs to make sure that the PC remains a cutting-edge device. The other reason is that Intel can't get away with charging $200 a chip for very long. Eventually its competitors catch up technologically and undersell Intel. Intel does then lower prices, of course. But it must also have its next-generation chip ready, to begin the cycle anew and keep margins high. Says Marc Schulman of Technology Strategies consulting firm in Stamford, Connecticut: "Intel is in the same position as auto manufacturers who want us to buy a new car every three years. It's planned obsolescence."

The business model allows little margin for error, which is why Intel can no longer afford to rely on others to create demand for PCs. Around 1990, Grove started moving the company from being an industry follower to a leader. Intel's 486 chip had started to catch on, but a potential problem was threatening to stall the next generation (later named Pentium): The speed of the microprocessor was starting to outpace the performance of the rest of the machine. The existing "bus," the internal network that directs electrons around the computer, would serve data at a rate far slower than the Pentium chip would be able to handle. Consumers might buy a PC to get the power of a Pentium and be disappointed.

Until then, bus designs had come from IBM or other PC makers. In 1990, however, Grove knew of no sufficiently fast designs in development. An Intel division had a proposal for a bus called PCI, but Grove didn't think Intel should get involved. "The notion that we would step forward and initiate a re-architecting of the computer for me was very strange," Grove recalls. "Where did we get off doing buses? I remember having a very heated argument with the executive who was pushing for it. But finally he convinced us that if we didn't do it, a new bus wouldn't become accepted." Today PCI is the standard bus on PCs.

The bus decision set the stage for what Grove calls his epiphany. At Comdex in 1991, Grove delivered a keynote speech involving what he referred to at the time as the "mother of all demos." He showed how a notebook PC equipped with PCI and special computer chips could receive E-mail messages and graphics delivered over a wireless network. At the time, that was a real breakthrough. Dell, IBM, and other companies endorsed the vision, and sent representatives to appear with Grove onstage. The demo was a smashing success. Grove was amazed at his company's ability to stage-manage so many players. He realized that this kind of leadership could become Intel's key competitive strength. "That was the 'Aha!' for me," he says.

Around that time, Grove visited his friend Steve Jobs, who had left Apple Computer to found Next (which Apple recently announced it would purchase). Next was building a high-performance but pricey computer specially tuned for multimedia applications and ease of use. Jobs demonstrated the machine for his friend, and Grove returned to Intel inspired. Says Frank Gill, an Intel executive vice president: "Andy came back and said, 'I want you guys to do that same kind of development for the entire industry, to make PCs as good as the Next computer.'"

Five years later, the fruits of that effort are visible:

THE LABORATORY. In 1991, Grove created the Intel Architecture Labs (IAL) in Hillsboro, Oregon, an R&D operation focused on the PC. The labs now house about 600 employees--mostly programmers--in two vast buildings set on flat farmland 25 miles west of Portland. Says director Craig Kinnie: "Our primary objective is to grow the market for all products, not just Intel products."

Many of the labs' projects involve creating new software. One project aims to make it easier for Websites to deliver video via the Internet. Intel programmers created something called the Streaming Media Viewer, which software makers can incorporate into their products so that consumers can view video as it arrives from the Web. For $199, any Website owner can buy an IAL-designed add-in circuit card that enables the site's computer to broadcast video. But guess what--the card will work properly only with a fast microprocessor.

IAL software is also helping popularize Internet telephony. Many small companies, like VocalTec and Quarterdeck, sell programs that let PC owners make long-distance voice calls via the Internet. For the most part, such products have been incompatible, limiting the number of people you could reach if you used one. IAL helped develop a new way to make calls, worked with the Internet industry to get it adopted as a standard, then gave away Intel software that met the standard. It also licensed the software to Microsoft, which gives it away on its Website. The moves forced VocalTec and the other telephony companies to modify their software to meet Intel's standard. Soon most Internet phones will be able to talk to each other. Crows Gill: "Until we took a role in driving the standards, Internet telephones were largely toys. Now there's an industry consensus around this specification."

Next Intel will promote software that it has designed for Internet videophones. Intel expects that manufacturers will build this function into PCs this year. If that happens, people who buy a high-powered PC this fall will get a feature that only three years ago required special hardware, like the videophone that AT&T sold for $1,000. That kind of innovation could persuade users to keep paying $1,500 or more every few years for a new PC.

GROVE IN HOLLYWOOD. On December 12, Grove & Co. made a glitzy Hollywood debut with the launch of the CAA Media Lab. A windowless, high-concept cave, the lab is housed in the Beverly Hills headquarters of the Creative Artists' Agency--Michael Ovitz's old hangout and still one of the most powerful forces in entertainment. The media lab is designed to show what the PC can do as a display medium and as a tool for the creation of new kinds of content. CAA plans to shuttle stars, musicians, directors, and other "talent" through the lab for tutorials.

At the gala opening, a couple of Intel execs chatted up Cheers veteran-turned-Larry Flynt impersonator Woody Harrelson. Later, in the lab's stylish mock living room, jazz musician Herbie Hancock and CEO of database maker Informix, Phil White, sat down together to watch an interactive digital promo starring Danny DeVito. DeVito conveyed Intel's seductive pitch--that technology now enables entertainers to communicate with consumers without a big studio as intermediary. "On the Net you can go direct to your fans," DeVito's digitized image told the talent from behind a digital cigar. "All kinds of media are getting married, and they're having a baby--a very lucrative baby."

Ron Whittier, Intel's point man on the media lab, is senior vice president in charge of content. He spends all his time figuring out ways to get compelling material--for entertainment, leisure, and business--ready for future generations of PCs. "The media and software industries have to be in sync with the platform," he says. "Otherwise it affects our time line for selling processors." Adds Avram Miller, Intel's director of business development: "I want people in Hollywood to understand that the time is now. The PC in the home will have higher production values--advanced graphics, higher resolution, and more interactivity--than any other platform." Television manufacturers, watch your wallets.

INTEL, VENTURE CAPITALIST. Intel now owns shares in over 50 companies, and the total market value of its investments exceeds half a billion dollars. Some of the companies in which it has a stake are the Palace, which creates virtual communities on the Internet; Digital Planet, which has developed episodic, interactive stories on the Web; and Willisville, a tiny outfit that is looking to combine shopping, chat, and storytelling online. Says Grove: "We make media investments in order to lure the companies into our infrastructure, into our world."

One of Intel's biggest media holdings is in CNET, which runs Internet sites and produces broadcast TV shows devoted to technology. Last July, just before CNET went public, Intel paid a little under $9 million for a 4.5% stake; it is now worth some $17 million. Intel helps CNET in a variety of ways. For instance, the companies are working together on Mediadome, a Website that made its debut on New Year's Eve. It aims to show off the most up-to-the-minute combinations of media and technology with such events as live Webcasts of technology-enhanced concerts. Intel brings in the talent, as well as movie and television properties that can be adapted into "Webisodes." "We work very, very closely with Intel," says CNET CEO Halsey Minor.

As a venture capitalist, Intel is aggressive, willing to spend on some pretty esoteric projects if there's a chance they might spark demand for the processing power only Intel can provide. Says an executive at a big PC company that also invests in new-media startups: "Intel gets there before us all the time in venture deals. You go, 'How did they do that?'"

One example: This year Grove embraced OZ Interactive, a small San Francisco outfit that specializes in 3-D Internet software. The firm defines itself in a press release as "scientists aspiring to art and artists hacking code." Grove is so impressed by OZ that he took Skuli Mogensen, the company's 28-year-old Icelandic president, to last summer's Allen & Co. media executive retreat at Sun Valley. Mogensen explains his product: "It's a browser that enables you to walk around in 3-D spaces and meet other people. You can create your own persona, change clothes, be a woman or a monster, or talk like a woman or a monster." Sound esoteric? Grove doesn't think so: He gave OZ early access to PCs equipped with Intel's new MMX multimedia technology, to ensure that OZ software was ready to roll when Intel introduced MMX in January.

Other PC development projects are even further removed from Intel's core activities. One example is systems integration. Intel is talking to Walt Disney about new ways of distributing its content electronically. It is designing an electronic videoconferencing network for the World Economic Forum, an influential group of CEOs and government leaders. And it is in the late stages of discussion with Starbucks about building a high-bandwidth network that would link together selected cafes. The companies are coy about providing details, but soon you may be able to send and receive video E-mail while sipping your latte at Starbucks.

Grove is particularly interested in pushing digital photography (see Digital Watch). "This is a big job for microprocessors," he says. "We want to take a lot of expense out of digital cameras and make picture taking less complex than it is today. We'll be working with Kodak, Konica, Sony, and Hewlett-Packard, among others." He's also focusing resources on Asia, Intel's fastest-growing market, to help adapt the PC to Asian needs. Intel has a 70-person software lab in Shanghai developing multimedia and 3-D content in Chinese.

Grove's PC efforts have already had a big impact. Michael Slater remembers that the PC of 1991 was "very weak" when compared with Apple's Macintosh. "But those differences have narrowed--very much because of Intel's attention to the platform," he says. Even competitors applaud Intel's market development. Says Jerry Sanders, CEO of AMD: "Anything they do that creates demand for computing power is a good thing. The whole industry is pulled along by that spending."

There's just one major complication--Intel is no longer completely in sync with Microsoft. Says Grove: "We are driven to move the PC platform forward with a single mind. Microsoft has a bit of a split interest. New computers matter to them, but the installed base matters as well." Microsoft makes lots of money supplying software designed for the 250 million machines sold in the past. While it also depends on success with new buyers, it does not depend on them as much as Intel.

Still, Microsoft and Intel do generally cooperate in trying to stimulate demand for PCs. The tension shows up when it comes to the Internet, where Netscape, not Microsoft, dominates the market for browser software. Says Intel's Kinnie: "If a Microsoft person thinks they have a way to use their technology to get a proprietary position on the Internet, they'll do it. But our overall motivation is always to grow the market. We see our role as establishing an open framework so everyone can go innovate their brains out." So when it comes to Internet software, Intel works as closely with Netscape as it does with Microsoft.

PC manufacturers by and large applaud Intel's efforts at innovation. They worry, however, that immensely profitable Intel might claim even more of their comparatively meager take. Last June, for example, MCI started selling, under its own brand, an Internet server designed and built by Intel. If MCI had not offered the server to its business customers, they might have turned to Compaq or another server company.

Such tension will only increase as Intel pushes the PC industry ever harder. But Grove doesn't think the company has any alternative. "I always worry that we're taking on too much," he says. "But we have a very crass reason to be a driving force. We build large factories and we have to fill them. It can't scare you when you don't have a choice."

Intel has no mandatory retirement age, and at 60, Grove has no plans to step down anytime soon. But in the January 13 announcement, he relinquished the company presidency to Barrett, who is 57. Though Intel won't confirm the widely held perception that Barrett is in line to succeed Grove, the move was significant. As he revealed in a cover story in Fortune last May, Grove has prostate cancer. He is no longer getting treatment, and the disease seems to be in check. Nevertheless, Grove wrote, "I know I will be stuck with this fear for the rest of my life."

Given the question of the CEO's health, it's fair to ask who is this man Grove has anointed. An outdoorsman, Barrett fishes on his private ranch in Montana, and resides in Phoenix because it's closer to mountains he loves to hike. He travels to Santa Clara three days a week, and works from an Intel office near his home the other two days. His wife, Barbara, is a lawyer who finished second in the 1994 Republican gubernatorial primary.

At Intel, Barrett is known as a no-nonsense manager who turned the company's manufacturing operations into a key strategic asset. They were barely adequate until he took them over in the mid-1980s. Since 1990 he has been chief operating officer. His role won't change when he adds the presidency on May 21--he'll still be in charge of the company's day-to-day operations.

Last May, in that FORTUNE article, Grove wrote, "I have a rule in my business: to see what can happen in the next ten years, look at what has happened in the last ten years." In the past ten years, Grove has turned Intel into perhaps the most self-reliant company in the technology world. In so doing, he has put his stamp on the company and emerged from the long shadows cast by Moore and Noyce. If Barrett someday takes over as CEO, it is Grove's legacy that he will promulgate. Asked where Intel is headed, Barrett says, "We picture ourselves going down the road at 120 miles per hour. Somewhere there's going to be a brick wall to cross, but our view is it's better to run into the wall than to anticipate it and stop short." Sounds an awful lot like Andy Grove.