(FORTUNE Magazine) – If machines can have souls, then maybe this is karma: Odds are that before the year is over, Dell Computer will dethrone Compaq Computer as the world's largest maker of PCs. It is a fitting turn of events. After all, Dell will only be doing to Compaq what Compaq did to IBM seven years ago.

Back in those ancient, pre-Web, pre-Napster, pre-Palm Pilot days, Compaq was a low-cost, nimble company focused entirely on selling so-called Wintel computers--PCs and servers that ran Microsoft's Windows operating system and Intel's microprocessors. IBM, on the other hand, was a high-cost, unwieldy company selling everything from PCs to mainframe computers. Today, Dell is a low-cost upstart focused entirely on the Wintel computer market, while Compaq is a large, relatively high-cost company that sells everything from PCs to multimillion-dollar supercomputers to consumer electronics gadgets like a $199 MP3 player.

The strange thing is that Compaq is exactly where it wants to be. It's not too bad a place, either.

During the past three years Compaq has tried to refashion itself as a mini-IBM, with some degree of success. PCs accounted for only 49% of the company's $42 billion in revenues last year, and 13% of its $569 million in profits. The rest came from servers, services, and storage products like disk drives. "Our strategy and IBM's are very similar," says Compaq CEO Michael Capellas.

In reinventing itself in IBM's image, Compaq is certainly taking a contrarian approach. These days the fastest-growing and most highly valued computer companies are focused on a single market--like EMC in storage, Sun Microsystems in Unix servers, and Dell. Building a broad, diversified computer company competing in all those realms, and more, seems to be a recipe for disappointing results. "For Compaq to grow at best-of-breed rates is tough," admits the 46-year-old Capellas.

But this may be the only path open to a company of Compaq's size, and the best insurance against bad times. Compaq is moving both upward and downward along the computer-industry food chain. By selling big iron, it gains more traction with business customers. By bringing out Internet appliances, it entrenches itself with consumers. As Capellas says, "When the market gets soft, we are better cushioned against a downturn because we are in so many different areas."

The problem is, by branching out in different directions, Compaq has multiplied the risks. Critics point out that not even IBM attempts to fight on so many fronts. In entering new markets, Compaq is butting heads with competitors that know their turf and will defend it ferociously. It isn't surprising, then, that while Compaq is outstripping IBM in revenue growth it lags far behind in profitability. Wall Street isn't sure Capellas can pull off this transformation either. The stock is down 11% since he became CEO in July 1999.

Still, the truth is that Compaq would have been fried if it hadn't diversified. The PC business is fast turning into the high-tech equivalent of the auto industry, with a glutted market, low profitability, and brutal competition. In their most recent quarters, Gate- way Computer, Apple Computer, and Compaq's consumer unit all lost money on PCs. Gateway ousted its CEO, and Dell, which has exacerbated matters by declaring a price war, just cut 1,700 jobs in its first-ever layoffs. While PC shipments have grown every year, the average selling price has been declining faster, so that revenues generated by box makers are expected to fall to $175 billion this year from $192 billion in 1997. Meanwhile, profit margins, once typically 20%, have shrunk to around 5% or even lower. "These days you're lucky if they're positive," says IDC analyst Roger Kay. In the stodgy auto industry, at least revenues are growing, albeit slowly, and most automakers are making money, with average profit margins in the 3.5% range.

What has happened to the once-roaring PC business? In the U.S., the market is simply becoming saturated, and the rest of the developed world is moving that way as well. "Virtually every business in the U.S. that needs a PC has one," says Kay. Making things worse, consumers and businesses alike have no driving need to upgrade their PCs. Faster bandwidth, not faster processors, is in demand. On top of these long-term trends is the economic slump, which has helped bring growth of PC shipments in the U.S. to a halt.

Compaq is facing an especially big squeeze because its cross-Texas rival, Dell, wants to keep on growing, no-growth market be damned. Dell boasts better margins than any other PC vendor, and Michael Dell is using that edge to launch a good old-fashioned price war. A few years ago Compaq might have followed Dell's lead to protect its position. Capellas won't, saying that he will cede market share to protect profits. But he can't afford to blow off $20 billion of revenues a year. So he has to hurry to make Compaq more efficient.

Capellas is starting out by cultivating his own garden--that is, Compaq's sprawling Houston headquarters, which is in the midst of a major overhaul. Until recently, nearly every computer that the firm sold to U.S. businesses was partially assembled at the home office; they were then shipped elsewhere to be customized with extra memory, disk drives, and other components. (Consumer PCs are assembled in Taiwan by other firms, and shipped directly to the dealer.)

That's all changing. Construction workers are busy converting one of the seven manufacturing buildings into office space--rolling rugs over concrete floors and installing cubicles where assembly lines used to stand. By next year, nearly all assembly will be off-site. And instead of starting to build a PC in one place and finishing it somewhere else, Compaq will handle both initial assembly and customization in one of four configuration centers, eliminating extra handling and cutting inventories. The company is doing the same in Europe.

Compaq is hoping these improvements will lead to Dell-like efficiency. Dell turns over its PC inventory 55 times each year, meaning that it has less than a week's worth of inventory on hand at any time. "We will be at 60 turns by end of year," boasts Michael Winkler, Compaq's 55-year-old executive VP of global business units. Last quarter, the figure was only 31, so he has his work cut out for him.

Even if Winkler makes his target, other measurements will take a while to rebound. Compaq expects its PC revenues to climb only 3% to 5% in 2001. Analysts are even more pessimistic. Richard Gardner of Salomon Smith Barney says, "We're projecting negative 2% to 4% growth"--analyst-speak for a decline. So Compaq's PC business is likely to get worse before it gets better, and will keep on losing market share to Dell.

Slipping in the marketplace means Capellas forgoes more than just PC revenues. He does not want Dell, or any other company, to horn in on Compaq's corporate accounts, where it could start selling PCs and end up stealing some of Compaq's more profitable server, storage, or service business. Capellas is also counting on technological leadership in the PC market to help Compaq vault into what he hopes will be the more lucrative post-PC era, when new Internet access appliances will unleash a wave of buying by businesses and consumers alike.

Capellas plans to take Compaq directly into the heart of the consumer electronics business, selling every sort of digital gadget that connects to the Internet. In a few years, that could include music players, videoplayers, telephones, computers, games, and electronic books. The lineup of Net devices will only grow longer as wireless technology becomes less expensive, more reliable, and ubiquitous.

Already Compaq has introduced a handful of new consumer appliances, all under the iPAQ brand. One of Capellas' favorite toys is his handheld iPAQ Pocket PC. Like other wireless PDAs, it's got the usual workplace stuff--his calendar, travel schedule, and e-mail. It also holds his favorite music. Ask Capellas to show you the device, and the first thing he's likely to do is have it play "Brand New Day" by Sting, or maybe Bruce Springsteen's "Born to Run."

Besides the Pocket PC, Compaq sells the Personal Audio Player, a pager-sized MP3 player designed for use during workouts or commuting; the BlackBerry wireless e-mail device; and the Home Internet Appliance, a slimmed-down, single-purpose desktop computer for surfing the Net. Compaq is also giving demos of a line of MP3 stereo equipment, which can hold a lifetime's worth of music.

Just how big could this business be for Compaq? "If you take the units, accessories, and some of the services that go along with it, it's a larger market in 2005 than the traditional PC market," predicts Winkler. So far the Pocket PC, which debuted in June 2000, has been the biggest hit. A typical model (with accessories like extra memory cards and wireless service) sells for about the same price as a standard desktop PC, around $1,000--but has close to double the profit margin. In the fourth quarter Compaq shipped more than 200,000 units, and it has 500,000 units on back order. In less than a year it has become a billion-dollar business. No wonder Capellas is excited.

Then again, so are a lot of people. Compaq may well be pursuing Internet appliances with more fervor than any other PC box maker, but plenty of other companies are lining up to take a crack at this market--Intel and Microsoft from the PC industry; Motorola, Nokia, and Ericsson from the cell phone business; Phillips, Thomson Multimedia, and, most formidably, Sony from the world of consumer electronics.

Compaq may have one important edge against these powerhouses: its mastery of the technologies that underlie this emerging market. Take wireless e-mail. Because the company develops both handheld devices that tap into the Internet and computer servers that power the application, Capellas argues that Compaq can develop better products than a company focused on one end of the equation. It's a compelling advantage at this early stage of the post-PC era, because the technical standards that will describe how a rainbow of different gadgets will hook up to different sorts of communications networks don't exist yet. As standards evolve, Compaq's edge will erode, giving way to companies like Sony that are more experienced at developing mass consumer electronics products. So Capellas is pressing on with urgency.

To understand why Capellas is fashioning this industry polymath, it's best to look at what he was handed when he took over. Former CEO Eckhard Pfeiffer had made two big purchases before the end of his reign. The first came in 1997, when he spent $3 billion to buy Tandem Computers (which runs most of the world's stock exchanges) in an effort to jump-start Compaq's sales of servers to large business customers. The next year he spent $9.6 billion to buy Digital Equipment, a maker of minicomputers. But Pfeiffer botched the mergers of these disparate businesses, and chairman Ben Rosen stepped in, replacing him with Capellas.

The choice at first seemed an unusual one. Capellas joined Compaq in 1998 as chief information officer, following brief stints at Oracle and SAP. But he had spent most of his career in the oil-services industry, working for 15 years at Schlumberger, where he held the post of CIO, among other positions.

When Capellas took over as CEO in July 1999, he recalls, "the entire employee base was demotivated." The company had split into three warring camps--Compaq, Digital, and Tandem. "The cultural clashes were ferocious," he says. The dot-com phenomenon was booming, too, making it even harder to retain employees. As a result, employee turnover was 18% a year, and even higher among top management. Products were late to market. The former Digital sales and service force, which Compaq had spent so much money and effort to acquire, was leaving in droves.

Capellas made a number of key changes right away. He replaced what remained of the executive team. Of the 11 senior staff who worked under Pfeiffer, only Michael Winkler remains. Capellas also reorganized the work force, mixing former Digital, Tandem, and Compaq employees. And he dropped businesses that were not central to the company, like routers. "Why try to compete with Cisco?" asks Capellas. "It is much better to partner with them."

Just as important, the gregarious and outgoing Capellas changed the mood at Compaq. Under Pfeiffer, an extremely reserved, formal man, the atmosphere was stifling. "You were underdressed if your cuff links didn't match your tie clip," says Capellas. He eliminated the dress code, and encouraged Compaq's 70,000 employees to have fun again. Take the elevator to the eighth floor early in the morning, and you're likely to hear music blasting from Capellas' office. He's also taken to playing a song before each quarterly conference call to energize his executive team and set a particular mood. Last time it was Cher's throbbing dance hit, "Believe."

But music and good feelings will only take Capellas so far. "Right now Compaq's still a turnaround story," says Ashok Kumar, an analyst with U.S. Bancorp Piper Jaffray. "The question is, What do you do for growth?" Compaq projects revenue growth of 6% to 8% in 2001, on top of 10% growth in 2000--"which is almost no growth," says Kumar.

The anemic PC business is only part of the problem. The other half of Compaq, the $22 billion server, storage, and service business, collectively dubbed Enterprise, isn't growing very rapidly either. Take the 38,000-person services organization. It's second in size only to IBM, with $7 billion in revenues. But sales declined 2% last year. Peter Blackmore, 53, executive VP of worldwide sales and service, says the group's problems have been fixed, but predicts that service revenues will grow only 6% to 8% in 2001. Another slow-growth business comprises the old Tandem and Digital minicomputers that run Unix and various proprietary operating systems. Revenues for this unit were flat in 2000, at $3.2 billion. New customers aren't buying proprietary computers these days, and in the fast-growing Unix market, Compaq is a distant fourth, trailing Sun, Hewlett-Packard, and IBM. Just as bad, storage products grew a paltry 3% in 2000, to $5.2 billion. That's a weak showing in a booming market. Revenues of EMC, the leader in this business, grew 32% to $8.9 billion last year.

Compaq can't count on quite as much growth from its strongest high-end play either. Wintel servers, Compaq's fastest-growing Enterprise segment, brought in revenues of $5.8 billion last year, up 27% over 1999. Compaq dominates this sector, with a 34% market share. But the growth rate is slowing as the business gets larger and competition from Dell and others heats up.

Overall, Capellas has done a fine job of righting a listing ship. He's made the company more efficient, boosted profits, and re-energized the employees. What he hasn't been able to do is get Compaq growing at anywhere near the rate it used to. He probably never will. Of the 15 companies with at least $42 billion in revenues on last year's FORTUNE 500 list, only six grew more than 10% from continuing operations, and none as much as 20%.

That may not be all that bad--if Capellas can figure out how to get more from what he has. IBM CEO Louis Gerstner has shown that a computer company can be successful even if it doesn't grow very quickly. In 2000, IBM's revenues were up a scant 1%. By that measure, Compaq's 10% revenue growth for the year looks fabulous. But IBM has one thing Compaq doesn't: solid profits, with operating margins of 13% last year. By contrast, Compaq's operating margins were 6%--which explains why Capellas' mantra for the company is "Drive Profitable Growth." The trick will be turning that mantra into karma.

FEEDBACK: enee@fortunemail.com