(FORTUNE Magazine) – You can tell things have changed at Microsoft just by paying a visit to Bill Gates. Middle age has set in--his face has lines; his hair, unkempt as ever, now shows a little bit of gray; you can see by the way he tilts his head that he's acclimating to his first pair of bifocals. But the real surprise is what's missing. Gone is the hollow look of perpetual weariness that comes with being the father of young children, the target of an historic antitrust case, and the world's best-known CEO. Instead Bill is all smiles and--dare we say--almost bubbly as he paces his blandly furnished office and describes how Microsoft's developers are creating cutting-edge technology as never before.

What's apparent is that work, for Gates, is fun again. Far from disengaging from Microsoft since his decision early last year to cede the CEO-ship to old buddy Steve Ballmer and appoint himself chief software architect, Gates has never been more absorbed. He now spends most of his time reveling in the details of Microsoft's product and technology strategies. The change seems to be paying off, and not merely for Bill's state of mind: A new cohesion of purpose and sense of inventiveness pervade the company.

This surge of vitality--which will become obvious this year as Microsoft rolls out an unprecedented wave of products and services--is sure to confound a lot of people. Just a year ago it was easy to imagine Microsoft as over the hill. It had lost its epic antitrust battle against the Department of Justice, a humiliating defeat that culminated in a court order, now under appeal, that the company be split in two. Microsoft baiting was in vogue--cartoonist Garry Trudeau even exposed a fictional employee as Satan in a series of Doonesbury strips. Customers, long resigned to Microsoft's practice of constantly upgrading its software and charging them over and over for the same program, grew cynical and rebellious, trading tips on how to save money by skipping generations of Microsoft products.

Worse, for all of Microsoft's stupefying success in spearheading the PC revolution, the Internet seemed to be eluding its grasp. Cisco, EMC, Oracle, and Sun--the so-called four horsemen of the Internet--displaced it as Wall Street's high-tech darlings. Microsoft's financial results were never bad, but revenue growth slowed enough that as the Nasdaq soared in late 1999 and early 2000, Microsoft's stock stayed relatively flat. And as dot-coms boomed, Microsoft suffered its first-ever brain drain: Gates and his lieutenants watched in dismay as scores of talented employees cashed in their options and went to work for hot startups. No wonder that last year, when Gates changed jobs, some questioned if it wasn't just a face-saving way to step back from the business to devote himself to philanthropy or just go play.

Wrong. Not only is Gates fully engaged, but the clouds over Microsoft have begun to dissipate as well. So far this year, the stock is up 62% (the four horsemen, by contrast, have seen theirs sink an average of 34%). Profits are pouring in so fast that Microsoft's legendary cash hoard, already $30 billion, is growing by $1 billion a month (see box). Even more important, the court of appeals, whose Microsoft ruling is due shortly, will almost surely reject the breakup of the company ordered by trial judge Thomas Penfield Jackson, and throw the case back to the district court. Impact on Microsoft: lots more legal fees, but not much else.

Back before the trial, Gates' company was so feared and disdained in Silicon Valley that internal documents from Netscape referred to it as the "Beast from Redmond." Well, the beast is back. And for rivals who hoped that the rise of the Internet--even more than the antitrust case--had begun to blunt its claws, we have more bad news. Before, Microsoft used its PC operating system monopoly to crush rivals and capture markets. Now it has learned something critics said it never could: to be creative. That, more than anything, is what will let Microsoft thrive--and maybe rule--in the Internet jungle.

To appreciate the change, you have to understand that Microsoft has long been dissed by the digerati as a big fat copycat. Its coders didn't even write the original software responsible for its meteoric rise--Gates purchased what became MS/DOS from a rival Seattle programmer in 1981 for $50,000. Microsoft has more than once found itself defending its methods of software development in court: In the 1990s, Apple Computer unsuccessfully sued, claiming that Windows violated patents that defined the famously friendly Macintosh graphical "look and feel." And the federal antitrust complaint hinged in part on Microsoft's copying Netscape's Internet browser.

In each contest, Microsoft staunchly defended itself by citing its fundamental right to "innovate." Yet its modus operandi has most often been to replicate and try to improve upon others' successful software ideas, and then absorb them into its products. That's pretty much how it overthrew Lotus and WordPerfect with its Office productivity suite in the early 1990s, how it is attacking Oracle and IBM with its SQL database software today, how it hopes to displace Sun's Unix-based enterprise and Web server products with Windows 2000, and outdo Yahoo and AOL with its oft-reconfigured MSN online services. All these lucrative digital franchises were pioneered by others and grew so fast that Microsoft's hyper-competitive brain trust found it impossible to just sit by and watch.

Not all the initiatives will work, of course--at least not right away. Wags joke that no Microsoft product is much good until version 3.0, pointing to Windows as the prime example: The company sold ugly-duckling versions for years before triumphing in 1990 with Windows 3.0. Still, with its try-try-again determination, Microsoft has become an immensely profitable $25-billion-a-year colossus, arguably the most successful tech company ever. In 25 years, it has never had a bad quarter, and a single share of Microsoft stock purchased for $28 on the day of its IPO in 1986 would now be a 144-share holding worth more than $10,000.

All that said, even the cleverest copycatting won't make those shares redouble again. For sustained fast growth, Microsoft will have to pioneer new markets, as well as compete and win in sectors where its desktop dominance does not guarantee a free ride: in videogame consoles, with its multibillion-dollar Xbox machine bet; in PDAs, with its PocketPC architecture; and in cell phones and Internet appliances. But here's a note to investors: If you had the opportunity, as FORTUNE recently did, to walk the halls of Microsoft's development groups, you might well come back with your checkbook poised.

The inventiveness blossoming on the company's rambling, coniferous campus in Redmond, Wash., is truly something new. "We're doing some of our best work ever," declares Gates. "In the last couple of years, a lot of our work hasn't been that visible. The lawsuit dominated the news. Then there was the whole thing about dot-com companies knowing it all. Not that we planned it this way, but there's actually a wave of products coming that will show we are at the beginning of a new era distinct from the Internet era." Or as Microsoft strategist David Vaskevitch puts it, "We are about to hit the golden age of software. The company is reinventing itself--let's call it Microsoft 3.0--so we'll be ready."

Invent We Must

What to make of all this? Microsoft is getting more creative in large part because it has to. For one thing, it has nearly exhausted the inventory of clonable ideas that are big enough to generate material growth. Revenues in the fiscal year ended last June grew 16%, down from a whopping 36% a year earlier. Also, Microsoft has come to accept that in one crucial market--high-dollar enterprise software applications--competitors like SAP, PeopleSoft, and Siebel Systems are so firmly entrenched that it can't attack them head on. It thinks it can profit by creating new kinds of software to augment and draw data from their applications.

Moreover, Gates is right that a new era is aborning. As digital devices increasingly link up, doing more and more business automatically among themselves, the Internet is evolving beyond the World Wide Web. Before long, the browser will no longer be the sole tool for interaction on the network. Instead, just about any software application on your PC--and even on your cell phone or PDA--will tap directly into interactive services to make work and life easier to manage. The services will probably sell by subscription, but there is no business model to follow for providing them--so Microsoft must help invent it. That is the vision driving its ambitious Dot.Net and HailStorm initiatives to transform the Internet. Microsoft also needs new ways to price and market software to spur sales in the many markets it has begun to saturate.

Beyond that, CEO Ballmer believes the time is ripe to go after a vast, fragmented market that has long stymied the infotech industry: small business. Companies with fewer than 100 employees account for more than half the output of the U.S. economy--but the software on the market for them right now is a crazy quilt of specialized applications.

There are also internal reasons why Microsoft is becoming more fecund. For years it has pumped more than 15% of revenue into R&D--$3.7 billion last year alone. That's paying off in a growing stream of innovations: from a whizzy, Internet-based videoconferencing system called the Ring-Cam to voice and face recognition to arcane but important data-mining technology that will come into play as apps begin to troll corporate and Internet sources for what Gates calls "business intelligence." Finally there's the human factor: Microsoft's top managers and engineers aren't precocious kids anymore, but seasoned masters of the game. Which brings us to the real driving force behind the new creativity--the guys at the top.

Three Heads Are Better Than One

Bill Gates has always wielded tremendous personal influence over Microsoft, from his obsessive competitiveness to his physical mannerisms to his trademark adolescent slang. (The beatnik-era "cool" was an all-purpose favorite adjective of Gates and his minions decades ago, long before its usage became cool again in popular culture in the late 1990s.) But over the years, as Microsoft grew into a 40,000-person behemoth, and as Gates, by virtue of his extreme wealth and outspoken nature, became a leading figure on the world stage, he lost touch somewhat with the inner workings of the company. "We got into a phase where I wasn't able to integrate products and strategy," he explains. "We were growing so much.... But we weren't quite as optimal in some technical areas as might have been great. Why are we not stronger in management software that lets you tell what is working in a business and what is not? If we had had great strategy coordination like I'm trying to do now, we might have that technology today."

Gates decided that Microsoft's stable of mostly homegrown managers was ready for more responsibility. His No. 2, Ballmer, had worked over two decades all around the company, leading software development groups, dealing with key partners like IBM and Intel, calling the shots in marketing, and building a massive worldwide sales organization. Ballmer, in turn, had hired or nurtured top-notch managers who were itching for bigger challenges. It seemed time to rethink how the company was run. The solution Gates hit on came in two stages. First, in early 2000, he made Ballmer CEO. Then last fall, he and Ballmer agreed that if two heads were better than one at the top of the company, how about adding a third?

They had the perfect choice in Richard Belluzzo, a seasoned executive who had run Silicon Graphics and also had been on the short list to head the legendary Silicon Valley powerhouse Hewlett-Packard. As president, Belluzzo would manage sales, customer support, finance, and other operations--making Microsoft's trains run on time--while Ballmer would focus on strategy, marketing, and being the company's public face. That would free Bill for software architecting. Suddenly Microsoft had three world-class CEOs, each concentrating on the disciplines he was best suited for.

"We hadn't thought that we could have our cake and eat it too, and, well, now that we are in that configuration, we say to ourselves, 'Gee, this is great,'" Gates chortles. "Maybe we should have done it sooner." He admits, though, that adjusting to the new roles took him and Ballmer months. "There was definitely some groping about to figure out who is supposed to be the good guy, who is supposed to be the bad guy when dealing with this or that group that is not doing an ideal job, but we love them anyway." (Can you imagine Gates as Mr. Good Guy? After our interview, he headed off to smooth the feathers of a product manager whose project had been reassigned to a different group.) For his part, Ballmer seems to thrive on being Mr. Outside. Sounding entirely CEO-like, he explains why the dot-com collapse plays to Microsoft's strengths: "We don't have all of the commotion and instability caused by random venture capital outfits soliciting our people for jobs that sound like get-rich-quick schemes. That, and having the stock market settle down, is good from our perspective, because it reminds everybody this is a long-term game. It's about having vision, being patient, executing, all that good stuff."

Managers up and down the ranks, especially in the product development groups, say they're glad to have Gates paying more attention, and in some cases acting as referee. Says group vice president Jim Allchin: "There are always conflicts among development groups, and two years ago Bill used to let us just fight it out. Now he spends time looking for technical synergies and driving us to achieve them." Bob Muglia, group vice president for personal services, agrees: "Not only is there less duplication, we're taking a longer-term view. Bill helps us think more broadly, rather than obsessing on our little piece of things."

"Creative" at Microsoft is not the same as "laid back." Gates is no less demanding as chief architect than he was as CEO, and his technical planning sessions are intense. The product development groups shower him with meticulously researched "opportunity maps" showing the competitive landscape for a certain software market; with detailed "scenarios" demonstrating how new software might be used in the real world; and with Microsoft's equivalent of management-consulting reports, which identify universal "schemas"--business processes within and between companies that are ripe for sophisticated automation. In short, Microsoft has become wonkier than ever, just like the boss.

The Ultimate Office

The best place to observe Microsoft's creativity in action is inside its rapidly evolving Productivity and Business Services group, one of the three major chunks of the company. While much of the innovation at the other groups would make sense only to nerds, nearly all of us can relate to the word processing, presentation graphics, and number crunching performed by Microsoft Office, the group's primary product. Jeff Raikes, a 20-year veteran who on the side co-owns the Seattle Mariners, is charged with expanding the Office market, as well as with Microsoft's campaign to wire small businesses and the effort to create software that can tap into SAP and other corporate applications.

Microsoft has a lot riding on Office. At an estimated $9 billion in annual sales and well over $4 billion in operating profits, it is the company's biggest moneymaker. The challenges it faces are a microcosm of Microsoft's--demand growth is flattening. For Raikes, a Nebraska native who spent the past five years leading the company's worldwide sales and marketing organization, overseeing Office is like coming home. It was he who, in the early 1990s, hit upon the strategy of bundling Microsoft Word, Excel, and PowerPoint into Office, a marketing masterstroke that gave Microsoft a cash cow in a league with Windows itself. Now, explains Raikes, "the challenge isn't to simply increase market share per se, because Office is already dominant, but, instead, as Jack Welch has said, to expand our definition of the market. We are trying to come up with features for Office that result in people spending more time using it each day--expanding our time-share, so to speak."

Raikes assigned a hot-shot MBA to figure out everything knowledge workers conceivably could do with a computer in a typical day. After several months' research, David Cumberbatch, a burly, soft-spoken 33-year-old Trinidadian, came back with an opportunity map of hundreds of activities--writing, numerical analysis, diagramming, multimedia presentations, e-mail, project managing, training, virtual conferencing, and on and on--divided into seven categories and projected to represent a $50-billion-a-year market by 2003. Spreading out the map on his desk for FORTUNE recently, Cumberbatch launched his explanation with a phrase worthy of that Microsoft character in Doonesbury: "Think of yourself as Napoleon..."

Interestingly, Raikes' team used the map not to plan market conquests but to hatch an idea. What if, in much the same way as a browser lets you view the content of a Website, Office could be modified to serve as a tool to tap into data in large-scale enterprise applications, making them easy to manipulate on the desktop? Raikes' group even came up with a term to describe this capability: the digital dashboard.

It sounds exciting, but it is hard to do. For a digital dashboard to work, Microsoft must ensure that Office can understand and accurately manipulate the data formats of hundreds of programs, which means winning the cooperation of enterprise software makers. Amazingly, SAP, Siebel, and dozens of others bought in--they earn little money on the desktop portions of their product lines and like the idea of making their systems more accessible. Microsoft also has to supply server software to act as an intermediary between the desktop and the enterprise systems. But Raikes' group was already working on something called the SharePoint Portal Server, which, with a few modifications, could do the job. Voila, Office had a new technology strategy and a broader potential market.

As FORTUNE went to press, OfficeXP, the first version to reflect the new approach, was set to debut on May 31. Like previous upgrades, it includes plenty of bells and whistles--new fonts, style sheets, document templates, and the like. But the digital dashboard feature amounts to a qualitative--and creative--shift. For Microsoft, it also marks a pivotal first step away from the upgrade treadmill. Ultimately, Microsoft hopes to sell Office on a subscription basis, which might include special charges for data services delivered via Office. But the company has put off rolling that out in the U.S. for at least a year.

Raikes' other big project is small business, and he's attacking on several fronts. First there's bCentral, Microsoft's Website and e-commerce hosting service for small businesses. Kathleen Hebert, the vice president in charge, explains, "For $24.99 a month, we'll host a Website and provide e-mail services, as well as a shopping-cart setup for e-commerce transactions, credit card clearing, and customer management." The idea of buying software as a service makes perfect sense for small businesses, she says: "They don't have the money, and more importantly the time or people, to build their own systems."

bCentral is just the beginning. Last year Raikes also engineered the $1.1 billion acquisition of Great Plains Corp., the leading maker of enterprise management and accounting software for small and medium-sized businesses. Microsoft will waste little time tuning the products to link up with OfficeXP. It is Microsoft's biggest outright acquisition to date.

Perhaps the most far-reaching project in Raikes' group could be thought of as the business equivalent of the human genome project. Norm Judah, a vice president who once managed Microsoft's computer infrastructure, leads a team that is trying to develop a schematic of every conceivable activity and interaction that any business might require, both internally and with customers and suppliers. Many of these activities already have been automated in piecemeal fashion by older data-processing systems at individual companies, but no one has ever tried to map out and standardize them all.

"It turns out that a lot of these thousands of business processes coalesce around very similar documents and records," says Judah. He has created a 70-square-foot "module map" to illustrate the flow of transactions and interactions. "The general ledger for one business isn't all that different from another. SKUs and UPC numbers are standardized. Even a power-of-attorney document can be reduced to a standard form."

The goal is twofold. Streamlining and standardizing electronic record-keeping and routine business activities would be a boon for companies large and small, and would make it easier for Microsoft to adapt OfficeXP to be the front end for even more business processes. It also would open the way for other "business intelligence" programs and services to help track, as Gates puts it, "what really works and what doesn't."

Though Judah's project is still a long way from completion, his boss, David Vaskevitch, sees it as the linchpin in Microsoft's strategy to lead the next transformation of IT. Says he: "When you really think about it, productivity programs like Office fundamentally changed the way people write and communicate in documents. If we're successful at this, we'll fundamentally change the way people interact with the economy."

Gotta-Have Gadgets and Software Everywhere

Several months before Gates quit being CEO, he and Ballmer concluded that Microsoft's "company vision"--a computer on every desk and in every home--was obsolete, mainly because it was so close to coming true. So they scratched their heads and came up with a new one: Empower people through great software--any time, any place, and on any device. The general reaction among Microsoft watchers was "Huh?" It seemed to include everything and nothing.

The clunkily worded vision finally is beginning to come into focus. Earlier this year, Gates and group VP Muglia introduced HailStorm, the first of Microsoft's "personal Web services" and an opening gambit in its Dot.Net strategy to upgrade the Internet to be more versatile and interactive. HailStorm's initial service, called Passport, provides an online repository for all sorts of personal information and privileges that you can tap into from any computer with a Web browser: contacts, credit card accounts, calendar, computer appearance preferences, file space for documents, clearance to use online versions of productivity applications, an electronic ID card, and more. It will give you access to your important information from anywhere and also simplify online transactions such as purchasing merchandise or airline tickets. Because HailStorm knows you already, no matter which Web merchant you deal with, it promises to let you transact your business with far fewer clicks and much tighter security and privacy.

That's one example of making software available anytime and anywhere. What about "on any device"? They're working on that too. Of the scores of development projects at Microsoft, Gates' favorite is a wireless gadget called the Tablet PC. As the name suggests, it's a portable, perpetually networked PC about the size of a short legal pad. You interact with it using a special digital pen, although you can plug in a keyboard if you want to compose a lengthy document. Microsoft will rely on partners like Dell, Compaq, and Hewlett-Packard to build it starting early next year, but the basic design has been concocted in Redmond.

Also slated for release later this year or early next is the Stinger, a cell phone that will incorporate the functions of a PDA--address book, calendar, audio and video capabilities, and Internet connectivity to give access to HailStorm services, e-mail, and Web browsing. It's Microsoft's attempt to outflank Palm, the PDA leader, which is relying on partners to integrate phones with its devices.

There's one other new product this year that will, in a sense, bring Microsoft full circle. In October the company plans to ship WindowsXP, a slick new version of its flagship product for office and home PCs that it hopes will displace the older Windows 95 and Windows 98 operating systems. WindowsXP is noteworthy not only because it begins to fulfill some of the promises of the Dot.Net strategy, but also because it eliminates the last vestiges of MS/DOS--the program Gates bought back in 1981--from the code, once and for all.

All this makes Bill happy. And when Bill's happy, Steve is happy. And when Bill and Steve are happy, the whole company seems to hum. Indeed, the best thing about Microsoft's new inventiveness may be that the process is so all-consuming it takes everyone's mind off the lingering shadow of Judge Thomas Penfield Jackson's breakup order.

Ballmer and Gates are so confident Microsoft won't be split up that they have no breakup plan in place. Even so, every Tuesday and Friday morning, wherever they are, Bill and Steve--and Microsoft's legal department and PR team--hold their breath until the stroke of noon Eastern Standard Time, because Tuesdays and Fridays are the mornings when the federal appeals court in Washington announces its decisions. But then it's back to work.

REPORTER ASSOCIATE Jessica Sung

FEEDBACK: bschlender@fortunemail.com