Scott G. McNealy looks mighty calm for a man running a company whose stock has cratered.

In the past two years, shares in Sun Microsystems have plunged from $64 to a lowly $3.28. A cool $195.1 billion in market cap has evaporated. Subtract Sun’s cash from the equation, and investors value the company at barely $1.63 per share, less than they’d pay for a slice of pizza. Now, from the coffee bars to the research and development labs of Silicon Valley, the buzz is that Sun, long a symbol of ingenuity and dynamism, is looking more and more like a relic of a free-spending era that’s long gone.

McNealy, Sun’s chairman and chief executive, says he’s not fazed. Throughout Sun’s 20-year history, he has grappled with crises before, struggling to convince skeptical investors, customers, and employees that Sun could transform itself. Each time, he pulled it off. In the early 1990s, tech pundits said Sun should ditch its workstation business and jump on the cheap Intel-Microsoft Windows bandwagon. McNealy wouldn’t listen. He went the other direction and built bigger and more powerful machines, setting up Sun to take advantage of the Internet boom. These days, the 48-year-old CEO, dressed in his trademark jeans and sneakers, says matter-of-factly that he understands why investors are treating his stock so poorly. "We’re not making money."

This year, Sun has shaved $600 million in costs from its supply chain software Honcho Schwartz (left) and consulting chief Sueltz

Worse, sales have taken a nosedive, down 32%, to $12.5 billion, from a high of $18.3 billion two years ago, as net losses over the past five quarters, excluding special charges, have mounted to $307 million. In the past two years, gross profit margins have skidded by 20%. Top managers, have jumped ship. And yet McNealy clings tightly to the formula that has worked for him before. It calls for stubbornness, hard work, and faith that the research and innovation that have kept Sun at the head of the industry through the years will come to the rescue.

There’s no time to waste. A fearsome posse of competitors, from Dell Computer to Microsoft and Intel, is battering its way into Sun’s core market for computer servers, selling low-cost machines at a fraction of Sun’s price. A few years ago, servers powered by Microsoft Windows software and Intel chips couldn’t perform in the same league with Sun. Now they can.

Worse, Linux’ open-source software is making inroads into McNealy’s market. It’s created by legions of volunteers, and it’s free—a price that’s hard to beat. McNealy finds himself selling the tech equivalent of a Mercedes in a market of Honda buyers.

Try saying that to McNealy. He maintains there’s a home for Sun at the very top of the industry, safely above the Linux- and Microsoft-powered hoi polloi. He fiercely resists the notion that Sun’s sophisticated servers could ever follow the brutal course of a commodity market. But if those $4,000 boxes rolling off the assembly lines at Dell Computer Corp. aren’t technically commodities, they behave very much like them, pummeling prices in Sun’s core business.

McNealy is battling disposable computers. And even when he looks away from the cheap Dells, he finds little relief. In the pricey side of the business, IBM, with its horde of consultants, is swooping into Corporate America offering the ultimate in no-headache computing: It will take over the entire burden of running corporate computer systems for clients. Says Microsoft Corp. Chairman William H. Gates III: "In terms of products that meet the market’s needs, [McNealy’s] in tough shape."

Yet McNealy has a plan, one that he says will lift Sun not only back to profits but to the apex of the Information Economy. At the heart of the plan is Sun’s classic franchise: heavy research and top-of-the-line computer systems. In a world of specialty players, Sun is a rare bird that designs its own chips and writes its own server software and computer chips. And McNealy’s sticking to his integrated model. He’s pouring research dollars into network software. His goal, stunningly ambitious, is to have Sun servers and Sun software running superefficient networks of the future—marvels that run virtually free of human attention.

In other words, while investors worry about Sun’s very survival, McNealy, ever the contrarian, is plotting a path to supremacy.

Sun’s Eldorado Sun is staking its software future on a vast project to automate the work in complex data networks. In September, Sun unveiled its nascent effort, called N1. Here’s what it aims to do by mid-decade: One System: Think of N1 as one huge operating system for the network. The software will automatically manage all of the computers, storage, and switches, making them work together as one giant machine. Virtual Resources: The software views all the tech gear in a cooperate data center as one virtual pool of resources. IF a big project taps out the supercomputer, N1 will automatically route the work to other computers. Low-cost Computing: By running computers close to capacity, the system will reduce the need for new machines. The software also will manage and update software installations. That will eliminate many of the tasks now handled by squadrons of systems administrators.

And he has certain strengths to build on. By hacking costs, McNealy has stanched much of the bleeding and says Sun will break even by next year’s second quarter. Net income has taken a beating, but Sun has generated cash in every quarter of the downturn. In the most recent quarter, Sun also bought back $500 million worth of stock and paid off $200 million in debt, bringing debt levels down to $1.5 billion. And he’s spending $70 million on a worldwide ad blitz.

But the clock’s ticking. Analysts say McNealy has only two years before low-end technologies in operating systems and chips catch up to his own. That’s precious little time to defend the company from the onslaught—and to broaden his high-margin beachhead in software. It’s fears about Sun’s business model that are giving investors the willies. And financial worries are on the rise. McNealy recognizes that hardware is unlikely to produce fat boom-level profits again. He’s hoping that higher-margin software, which now makes up only 5% of Sun’s revenue and an estimated 9% of profits, will pick up the slack. He won’t predict when, but a bullish report from Merrill Lynch & Co. says that within two years Sun could generate up to 9% of revenues from software. To reach that target, McNealy needs faultless execution and more than a little luck. He must come up with new network software that matches the best in the business, from IBM’s to Microsoft’s. It’s a tough challenge for a box maker. Indeed, McNealy can only hope the company learned from its failure in the late ’90s when it treated software as an ugly stepchild to its booming server business—and blew a golden chance to run away with the market for e-business software.

McNealy must also plow into services, but without disrupting relationships with consulting partners, such as Electronic Data Systems Corp., that install Sun systems. Perhaps most difficult, he must convince workers, many of whom hold stock options that are deep under water, to bust their gut for a company many in the tech world are writing off. And they must hurry: If the cheap servers climbing up the food chain catch up to Sun’s top line before McNealy’s plan has traction, he’s in trouble.

As McNealy leads his company up this steep slope, he’ll doubtless face some tough choices. He’ll likely be forced to let loose a few of his precious technologies. High on the list are Sun’s proprietary Sparc chips and Solaris server software, which together eat up more than $200 million of R&D investment annually, according to analysts. But even here McNealy faces a dilemma. First, he has lots of customers who rely on maintenance and upgrades for these proprietary components. Cutting back R&D could cripple an important source of revenue. What’s more, if he ditches the very pieces that make Sun special, he runs the risk of tumbling into the cutthroat commodity world below.

McNealy’s Plan for Reigniting Sun Follow the Money: Software products generate gross margins around 80%, twice that of servers. McNealy has dedicated 1000 sales people to software. He’s also spending $900 million on it, half Sun’s R&D budget. Learn to Love Linux: Customers are clamoring for Linux, the free operating software. In the fall, Sun rolled out Linux on its low-end servers. With time, Linux will run on the powerful boxes too. Take on Cut-throat Rivals: In a bid to grab 30% of the market for supercheap Linux servers in a couple of years, McNealy has unveiled new servers priced as low as $2700. He’s using Intel chips and outsourcing production in an effort to still make money at that price. Beef up Services: Sun has doubled its force of consultants, to 13000, over the past three years, getting 32% of revenues from services. Longer term, he hopes to offer more lucrative consulting. Push Innovation: McNealy is betting the farm that his outsize R&D budget will help Sun become the first company to provide software for smart networks that look after themselves-without calling on costly human help. Sun has a hoard of $5.2 bn in cash and securities

Will McNealy hoist Sun back to the top? More likely is a Sun that settles into a specialty niche, providing high-margin servers with all the bells and whistles built in, a path similar to the one trod by Apple Computer Inc. in the consumer market. A drearier possibility: Sun could follow the footsteps of Digital Equipment Co., which failed to keep up with cost-saving changes in the computer industry and was eventually bought by Compaq Computer Corp. in 1998 for $9.6 billion.

If McNealy comes up short, the effects will be felt far and wide. The ideas pouring out of Sun’s labs have made the midsize computer maker into an outsize thought leader. Indeed, Sun has been strong enough to take on mighty Microsoft. And many in the computer industry maintain that integrated manufacturers, like Sun and Apple, which focus on entire systems, generate far more creativity than the component-based champions such as Microsoft, Intel, and Dell. Yet creativity doesn’t always win the day. And as Sun struggles, Microsoft grows stronger. Still, McNealy has long enjoyed a parallel career as the industry’s anti-Microsoft ringleader. But now he barely has time for Bill-baiting. He has a company to rescue. Friends say it’s times like these that get McNealy fired up.

McNealy is working to put in place the management controls and succession planning learned at the side of the man he considers his mentor, former General Electric Co. Chairman and CEO Jack Welch. McNealy has cut a layer of management that kept him from execs on the front lines. He’s acting like a battlefield sergeant, making sure his managers are following his strategy for facing the low-cost onslaught. When he’s not meeting with customers, McNealy is making needed repairs to his company. In July, he created an executive vice-president position for software for the first time, organized all of Sun’s consulting under one person, and assigned 1,000 salespeople to hawk Sun software.

But McNealy’s first big management changes didn’t come off well. About 18 months ago, he started work on a succession plan patterned after what he had learned while sitting on the GE board. Just like Welch, he planned to evaluate and train a new generation of leaders. At an April, 2002, powwow for Sun’s 200 vice-presidents, he unveiled his plan. Then it blew up in his face. Coupled with the downturn in Sun’s business, it looked as though the top people were jumping ship—or McNealy was forcing them out. Sun’s stock dropped 10% the week after Zander announced he was leaving. McNealy was so flabbergasted by the debacle he called Welch to ask what he did wrong. "Jack said, `Don’t worry about it. They’ll forget about it soon enough,"’ says McNealy.

To kick-start business in the low end, McNealy is making a tactical retreat in servers. He has opened up a place for the commodity components, Linux and Intel chips, in his economy offerings. To proclaim his newfound love of Linux, McNealy showed up at a Feb. 7 conference in San Francisco in the costume of a penguin, the Linux mascot. The challenge for McNealy is to crack open a door to low-cost business without encouraging high-end customers to swarm through and switch to the cheaper fare.

McNealy has aggressive goals for the Linux servers. He’s not airing them publicly, but several Sun executives say McNealy has told them that within the next two years he wants to grab 30% of what will then be a $6.5 billion market, according to researcher IDC. To take on cutthroat rivals, McNealy is keeping costs low and outsourcing production of Sun’s Linux machines, which start at $2,700. That’s a highly ambitious goal, given that Sun has no market share today. But if McNealy can hit his target, the low-end gear could add $400 million in gross profits in 2004.

That may be just enough to cover the slide in sales of Sun’s midrange Solaris machines. The trouble with such calculations is that Dell and its low-cost collaborators represent a fast-moving target—one that slashes prices to woo new business. Merrill Lynch estimates that from 2001 to 2004, revenue in Sun’s midrange will fall 60%, to $1.7 billion. That translates into an estimated $500 million hit on profits. McNealy’s bid hinges on his ability to focus the company on a handful of key initiatives.

One is services. As Sun extends its business from the boxes to the broader network, it will be up to the service staff to help customers install the full gamut of Sun offerings. Sun has 13,000 service consultants, double three years ago.

And while they’re dwarfed by IBM’s 180,000 consultants and HP’s 65,000-member force, they appear to be gaining traction. In the most recent quarter ended Sept. 30, service revenues were up 9%, to $879 million. Patricia C. Sueltz, Sun’s executive vice-president in charge of consulting, targets $1 billion in quarterly services revenue by next year.

Equally vital for McNealy is software. While Sun has been a bold software innovator, coming up with such advances as Java, software has remained a niche business marked by Sun’s failure to turn leading-edge technology into sales and profits. Now McNealy needs it more than ever. Software is at the center of his vision of a Sun-driven networked world—and it delivers gross profit margins of 80%. It’s little surprise that McNealy is plowing more than half of his R&D budget into software, much of it for Web applications.

Again, Sun risks new battles with old friends as the company plunges into different businesses. For years, Net software maker BEA Systems Inc. was an enthusiastic Sun partner. It created software to run on Sun’s Solaris servers that let customers deliver applications effortlessly via the Web. Now, Sun’s assault on the Web-software business has pushed BEA into the arms of chip giant Intel.

At the heart of McNealy’s vision is an ambitious software project called N1. Sun’s software developers have been working on the technology for two years, tucked away in a space-age data center at Sun’s Sunnyvale (Calif.) facility. The idea is to create vast networks in which the software administers itself. If one computer runs out of memory, the software seeks spare capacity elsewhere on the network. If the software develops a glitch, the program itself will work to fix it, without calling on costly human administrators. Sun will be releasing the first components of the program by the end of the year.

The trouble is, McNealy must invest heavily in N1 just to stay in step with competitors. IBM and HP are hard at work on very similar systems.

With all these challenges, it might make sense for McNealy to shelve his ongoing war with Microsoft. But he has trouble letting it rest. McNealy’s colleagues urge him to focus on more pressing threats. Before a Sept. 18 speech to Sun customers at San Francisco’s Moscone Center, his vice-president for software, Jonathan Schwartz, bet his boss $2 that he couldn’t avoid mentioning Microsoft during his speech. McNealy took the bet—and collected his money after his talk.

At the event, Sun’s first big customer conference in seven years, dreadlocked drummers on stage were pounding a beat when McNealy jumped up, beating on a drum of his own. He promptly launched into a 45-minute stump speech defending Sun, one of the last of the integrated computer makers. "There is no automobile-integration industry," he says. "You get a car fully assembled. They even wash it for you." Jokes and debating points aside, McNealy has to get Sun making money again. Only then will he convince the world that Sun can shine anew.

By Jim Kerstetter in Menlo Park, Calif., with Jay Greene in Seattle and bureau reports