THE MORE ROBBIE BACH THOUGHT about Microsoft entering the video game console scene, the more it made sense. "My opinion was that the company was a little stuck. The PC was going great, but Microsoft was struggling to figure out how to expand beyond the PC itself."

Microsoft had a solid track record in capitalizing on new technologies. Web browsers, operating systems, programming environments, consumer software, peripherals like joysticks and mice--the company carved out space in those areas. Video game consoles seemed the next logical step. "The PC market had started to flatten out. For me, this was an exploration into how do you drive growth in an adjacent market?"

On the DirectX team, Seamus Blackley had similar thoughts. He held no interest in things like market share or sales forecasts. But as he watched Sony brass announce the PlayStation 2 with great fanfare in early March, he knew everything they were saying was wrong.

In 1999, companies such as 3dfx and NVIDIA warred over the emerging PC accelerator market. Accelerators, also known as video cards or GPUs (graphics processing units), gave users dedicated hardware to render 3D graphics. Few games required them to run, but by the end of that year, id Software's Quake III: Arena would drive huge sales in accelerators.

Blackley was impressed by the technology, though he nursed a small grudge. "The advent of 3D accelerators was part of what caused Trespasser to look bad," he says. "It had this algorithmic renderer that could draw a forest, but it was completely unsuited to any of the first-generation graphics accelerators."

Now, in what he thought of as religious penance, his job was to test accelerators and convince their manufacturers to add support for DirectX to their games. That was why the PlayStation 2's announcement surprised and frustrated him. "Sony started making all these claims about the future of graphics: how they were going to replace the PC, and they'd run Linux on PlayStation 2, and they had the Emotion Engine. I got really frustrated. It seemed clear to me that Microsoft could wipe out Sony. I was also offended in the way nerds get offended by marketing bullshit."

As the person in charge of scrutinizing accelerators, he knew there was no way the PS2 could rival PC games. Consoles couldn't be upgraded. Within the next 12 to 18 months, Sony would have to lock down the PS2's design so the console could be mass-produced. Meanwhile, companies like NVIDIA and Intel would keep improving on 3D graphics and putting out more sophisticated cards. Developers made games for consoles, but they wrote those games on PCs.

Logically, the PC should be the undisputed, heavyweight champion of gaming. The only thing holding it back, and one of the problems DirectX sought to mitigate, was a lack of standardized hardware--a console's greatest weakness, but also its greatest strength. "You couldn't optimize the crap out of a game because you didn't know what kind of PC it was running on. I had just come from the PC games industry, and doing compatibility was a horrible problem."

On his flight back from visiting his girlfriend, Blackley had an epiphany. Since everybody used the PC to make games, and the roadmap for graphics tech on PC is spectacular and getting better practically by the day, the only thing holding developers back was the lack of a standard for a high-end platform. Microsoft could make a platform powered by DirectX and more than capable of destroying the PlayStation 2.

"If you're going to write a really great game--responsive, beautiful--you want to optimize it to within an inch of its life," Blackley says, "and it's impossible to do that if you don't know the exact hardware specifications you're operating on. You'd be developing on the target platforming you'd deliver to, which is not the case on any other game console; the development system is always different from the delivery system."

Blackley returned to Microsoft and wrote a memo that outlined his vision for going head-to-head with Sony. The first person he showed it to was Kevin Bachus. Like Blackley, Bachus developed an early interest in programming games. He deviated only once, graduated from the University of Southern California with a film degree. Finding Hollywood too irrational, he pivoted to tech journalist, then to marketing guru at game publisher/developer Mindscape, and finally to Microsoft where he leveraged his skills for the DirectX team as its product manager.

By the time Blackley came to him with his manifesto, Bachus had spent two frustrating years pushing Microsoft to do more in gaming only to realize the company would never market Windows as a game platform because it didn't have to. "The most important thing about DirectX was not only did we develop this technology that allowed you to develop high-performance applications in Windows, it also was the center of the all the thinking in the company about the game industry," Bachus says. "When Bill Gates or Steve Ballmer had a question about Sega, they would call us and we would talk about that stuff."

Bachus not only agreed with Blackley's plan, he had been thinking along the same lines. "Even before I went to Microsoft, a lot of us had said, 'It's really weird that Sony, the Walkman company, is going to make a game console.' But they were very successful, so we said, 'Who else might bring legitimacy to our little industry?' The obvious answer was Microsoft."

Blackley thought so, too, and had written as much in his memo. "There was a little footnote at the end," Bachus recalls, "and it said, 'Oh, by the way, we should totally build a custom version of Windows and a hardware reference design, and get the PC manufacturers to build something that would compete with Sony. This was almost a footnote: 'We should do this, but we probably never will.'

That Microsoft would produce a console seemed like a longshot. It was a software company. That led organically to the idea that Microsoft could leave a bigger footprint on tech by making appliances that run Windows other than PCs. That was fine, Blackley reasoned in his memo, because Microsoft wouldn't need to manufacture game consoles and/or other appliances itself. Any arrangement Microsoft made would function similarly to their partnerships with OEMs: Just as Dell and other companies manufactured PCs and bundled Microsoft software, third-party manufacturers would handle production of hardware while Microsoft supplied its DirectX-enabled platform.

Otto Berkes liked the initiative. A manager on the DirectX team overseeing development of the software's perpetual next version, he believed in Blackley's memo and went to talk with Ted Hase, head of evangelism for DirectX. The next day, Hase came to Bachus. "We're thinking of trying to sell this idea," Hase said. Bachus took Blackley's paper and converted it into a PowerPoint--"because you can't order lunch at Microsoft without making a PowerPoint," he says dryly-- to organize information and show how this venture might function as a business unit.

Over the next few months, Blackley, Hase, Bachus, and Berkes became known as the Four Musketeers. Hase, in high enough standing at Microsoft to have his own conference room, designated it a pseudo-official meeting spot when they wanted to talk about their console, which they designated the DirectXbox. The guys would sit around the round table and munch on Jelly Babies as they brainstormed. "We'd have these meetings where we'd talk about what the next step should be, who we could talk to, who we could connect to, which senior guy would listen to us or wouldn't," Blackley remembers.

The guys recruited by testing the waters with managers adjacent to DirectX. Nat Brown, a software engineer whose job entailed scouting opportunities for synergy across departments, threw in with the Musketeers and made introductions to others he thought would be sympathetic to their cause. "Nate was pretty instrumental in connecting us with old-school people," recalls Blackley. "Rick Rashid ran Microsoft Research. People who'd been around Microsoft a long time really appreciated our entrepreneurial spirit."

Inevitably, Brown made overtures to Ed Fries.

LIKE A PROUD PARENT, ED Fries had watched his games group undergo a tremendous growth spurt. "There were fewer than 100 full-time employees when I started," he says. "We probably had 500 or 600 working on games" by the spring of 1999.

The success of Age of Empires and Doom 95, and acquisitions such as FASA, had opened doors Microsoft needed to expand Fries's game-publishing ambitions. He was even thinking of mobilizing a larger sports initiative to compete against Electronic Arts, whose franchises like NBA Live and Madden NFL owned most of that market. Fries was aware of Microsoft's conspicuous absence in the console space, but he'd never thought seriously of making a move. "Getting your first five percent share in the console business looked a lot easier than getting five percent more of the PC business. We had built some franchises and had some other things coming."

When a DirectX evangelist named Seamus Blackley got in touch and asked to schedule a meeting, Fries wasn't sure what to expect. Blackley entered Fries's office one afternoon with Otto Berkes and Kevin Bachus, and made his pitch: Microsoft should partner with OEMs to build and sell a game console powered by DirectX.

To Blackley's memory, Fries responded by laughing.

"He actually laughed. I remember being so bummed out by that. So bummed out. In fact, I have to be honest, when I think of that even now, it bums me out. It makes me question myself."

Fries's memory is fuzzy on that detail. "I don't remember it well enough to say whether that's true. It could be true. I'm pretty sure I had no idea why they were coming to meet with me," he admits. If he did laugh, he believes the response would have been in reaction to the audaciousness of Blackley's notion, and his own knowledge of a similar ventured he believed fated to crash and burn.

In May 1998, Microsoft announced that it would customize Windows CE, a version of its operating system optimized for portable devices like phones and tablets, to Sega's Dreamcast console. The derivative OS, codenamed "Dragon," would integrate features from the forthcoming DirectX 8.0: Rather than power the Dreamcast, Dragon would spin up and feed relevant information such as graphics libraries when it detected software optimized for the OS. Sega touted the partnership by placing a Compatible with Windows CE sticker on the front corner of the Dreamcast hardware.

Bizarrely, Sega developed its own custom software and encouraged developers to adopt it. Ultimately, fewer than 50 games used Dragon. Although Dreamcast had yet to launch in the US when Blackley and the other console advocates paid him a visit, Fries predicted failure for the Microsoft/Sega partnership. "I hadn't been supportive of the Dreamcast work that had been done to that point. It seemed very half-assed, a way to put a Windows sticker on the Dreamcast box. But this was a more serious plan."

The more Blackley and the others talked, the more convinced Fries became. "It would be a PC in disguise running a shipping version of Windows, but it would have an extra layer that made it act like a game console," Fries explains. "You'd be able to read PC game discs. When you put them in, the way you insert a disc into a console, the console would install it in the background and run it from the hard disk in this machine."

If this console came to fruition in the way Blackley described, it might be possible for users to stick one of his PC games into the console's tray and play it with a controller. As a bonus, assisting in a console effort would be a great way to study that market.

Though stung by Fries's first reaction to his proposal, Blackley came to see the other man as a kindred spirit. "If you're a senior Microsoft executive in charge of a division, and you own a lot of stock, and you have a really big house with a bunch of land, and your own dock on a lake, like Ed has, you don't back an idea like this for practical reasons. When it's an idea from two goofballs you laugh at the first time you hear it, you don't back it for any reason other than that it's exciting."

Things happened quickly, more for some members of the nascent DirectXbox team than others. Members like Berkes, Hase, and Fries thought about it when they could spare time. They had day jobs running businesses within Microsoft. Blackley and Bachus threw all their energy into the project. "Seamus and I really didn't have a lot to lose, and I never believed for a second that this wouldn't work. So we left our jobs and went to work for Rick Thompson in hardware."

Advocates such as Hase, Berkes, and Nat Brown continued to point Bachus and Blackley toward individuals whose talents would be crucial to getting their project off the ground. J Allard, the passionate manager who had galvanized Bill Gates to go all-in on the Internet, joined after talking over the venture with Cameron Ferroni, an engineer on the Windows team. Todd Holmdahl, a veteran of the hardware department, expressed interest. So did Jennifer Booth, a product manager for PlayStation from 1993 to '95, at which time she joined Microsoft and gave the DirectXbox team advice on how to approach marketing a console in the States. Don Coyner, a marketing previously employed by Nintendo of America, appreciated their boldness and joined the brigade, but worried privately that the eager gamers underestimated the steepness of the road ahead.

"The emotional part of me said, 'Hell yeah. It would be a blast to launch a console again,'" remembers Coyner, who played a key role in launching the Game Boy and Super Nintendo. "From a practical point of view, I said, 'Ugh. This is going to cost a fortune. These guys have no idea how much this is going to cost.'"

Among all the other plates they juggled, Bachus and Blackley took every opportunity to reassure higher-ups that a console was in Microsoft's best interest. One afternoon, Bill Gates received an email from DirectX co-creator Alex St. John, who'd since left Microsoft but kept in touch. Gates forwarded the email to Bachus. "Bill got an email from Alex St. John saying, 'You guys are fucking up. PlayStation is going to wipe the floor with you. You should have listened to me,'" Bachus recalls. "Bill sends it to us and says, 'What do you think about this?'"

Bachus got the impression that Gates had forwarded the email while chuckling and shaking his head. Another instance of St. John being his usual brash, pompous self. "He actually raises some good points," Bachus wrote back. "And in fact, Bill, we even crafted circumstances to get feedback from the industry on building a console. This is exactly what we planned for."

That, Bachus clarifies, was a bald-faced lie. The trick to answering a BillG question, he knew, was to project confidence. "This sounds good," Gates wrote back. "Keep it up."

Bachus and Blackley exulted. "That's currency, man," Bachus says. "That was a gold bar we could put in front of people and say, 'Bill likes this and wants us to keep up the good work.'"

Elsewhere, another group of believers fought to earn Gates's seal of approval. "What really happened was it turned out that once we got some traction with the executives, the WebTV guys showed up," Blackley says.

IN 1997, MICROSOFT PURCHASED TECH startup WebTV for $425 million. Craig Mundie, the executive who spearheaded the acquisition, saw vast potential in the company's engineering talent. Not only would WebTV receive a place of honor within Microsoft as the company's first home appliance, Mundie thought the engineers behind the project had what it took to build a game console.

Mundie's read on WebTV's engineering team was spot on. Some members of the group, which included Nick Baker, Dave Riola, Steve Perlman, and Tim Bucher, had worked at The 3DO Company, a studio founded by Electronic Arts founder Trip Hawkins and purveyor of the titular 3DO Interactive Multiplayer console. Hawkins's vision for the 3DO was an all-in-one entertainment center. Significantly more powerful than the 16-bit Super Nintendo and Sega Genesis, both of which ran games from cartridges, 3DO boasted 32-bit processing and played games on CD-ROM. When users finished playing, they could pop in their favorite music discs.

All that processing power came at a high cost. Hawkins launched the 3DO Interactive Multiplayer in October 1993 for a staggering $699. Much of the cost was attributed to manufacturing. Lacking the resources to engineer and distribute the console, Hawkins had pitched hardware licenses to Sony and Panasonic. Sony had demurred, already putting together its PlayStation. Hawkins courted Panasonic and Sega, but the latter backed out as well because of concerns over costs. Panasonic became a licensee and slashed the price of its version of the console to $199 to spark interest. By July 1996, most developers had dropped support for the Interactive Multiplayer, including The 3DO Company.

Now at Microsoft, the ex-3DO developers saw an opportunity for redemption. With Mundie's support, they wrote documents describing a game console marketed similarly to WebTV: An appliance running on a version of Windows CE stripped of everything except the resources it would need to power a game console.

Over a series of meetings and email chains, the WebTV group sneered at the DirectX team. They knew nothing about game development or consoles, the WebTV said of their peers. They were especially ignorant of hardware; they didn't even know how to price screws. "I'm thinking to myself, The fuck? You can get tons of people who can price screws? Why would people trying to figure out a business initiative worry about that?" Seamus Blackley remembers. "What they were trying to do was gatekeeping, which is what people call it today."

Despite their incredulity, Blackley and Kevin Bachus bowed under the weight of the WebTV crew's accusations. Blackley and Bachus were nothing at Microsoft. The other guys ran a successful product and knew how to ship a game console, albeit one that had tanked. Blackley grew despondent. Not only did his word not carry any weight, he still viewed himself as an exile from the games industry, and an embarrassment following Jurassic Park: Trespasser's public failure.

He was also stung by the on-again, off-again interest of Ted Hase and Otto Berkes, both of whom conveniently dropped off the radar when the DirectXbox project seemed in danger of being mothballed. Only Bachus remained by his side, steadfast and enthusiastic through every accomplishment and setback. "You know how it goes," Blackley says. "You have a bad meeting, or somebody doesn't support you, or somebody said to somebody they thought this would never work. Suddenly people stop returning your phone calls and emails."

Bachus and Blackley redoubled their efforts in the face of the WebTV group. When they weren't eating Jelly Babies, evangelizing the DirectXbox within Microsoft, or mulling over business concerns like how to organize a business model, they were buying and reading business books. "It was me and Kevin," Blackley says of the console. "Me and Kevin spent so much time together. That's from whence it sprang, to be totally honest. There were many people involved, but that's where it came from."

Fries, however, was unimpressed by what the WebTV guys' plans. The DirectXbox held much more potential, and he was reticent to support any game console-related endeavors involving Windows CE.

By May 1999, it was clear that one group would have to go. Managers from both groups arranged for a showdown: a meeting, known as a beauty contest, with Bill Gates and Steve Ballmer in attendance, during which both teams would bring their best presentations. Fries, the longest tenured between both groups, was familiar with such showdowns. "It was a classic corporate meeting where two groups of people are fighting over territory and who gets to do what project," he says. Besides PowerPoint presentations and gusto, both groups rallied executives to back their claims at the meeting. Rick Thompson and Robbie Bach, who headed up the consumer division, attended to support the DirectXbox.

On the day of the beauty contest, Fries stared across the invisible line dividing the conference room at Jon DeVaan, an engineer he'd worked alongside during his days on the Excel team. Craig Mundie kicked things off on behalf of the WebTV group. Everyone knew Sony's PlayStation 2 posed a threat. The question was, would Microsoft counter with the DirectXbox, a PC in the guise of a console; or an appliance like the WebTV?

"What they pitched was very much a PlayStation-like--custom hardware, custom software--game console," Fries recalls.

Nat Brown, Blackley, and Bachus spoke next, with Ted Hase remaining silent; his main job, one he performed in meetings for all his projects, was to listen and take the temperature of the room by reading body language, then offer advice on how to proceed. Brown loaded a PowerPoint and explained that the DirectXbox--which the team had already begun referring to as Xbox--and explained that Xbox would take advantage of Microsoft's strength by running Windows and DirectX. Microsoft could hire manufacturers to build its Trojan horse, PC-disguised-as-a-console for a launch in the fall of 2000, in time to counter PlayStation 2.

Game developers would support the Xbox, the group continued, because it would use technology they already knew from writing PC games. The specs they planned to include in every box include a processor from Intel or AMD, an accelerator from NVIDIA or 3dfx, a modem along with a broadband port for high-speed Internet connectivity, and, most controversially, a hard drive. Including a hard drive would add substantial manufacturing costs, but as far as anyone knew, Sony had no plans to include one. Including a hard drive would make it so devs could install games to the Xbox, resulting in faster load times compared to continually streaming data from a CD or DVD. They estimated a total manufacturing cost of $303 per Xbox, which they could upgrade every few years, saving consumers the work of choosing PC parts and installing them themselves.

Across the table, the WebTV team said the Xbox would never work. The parts were too expensive, especially the hard drive, which the crew saw as unnecessary. Ed Fries disagreed. Including a broadband port would enable Xbox to support high-speed online gaming, and a hard drive would be necessary to store and access games quickly. As a bonus, developers could use the hard drive to save games, a move that would save consumers the cost of memory cards like those used on PlayStation. Gates listened carefully to Fries, who had earned his respect years ago for his work on Excel and Word.

The argument flowed back and forth. One of the WebTV developers brought up a salient point: A Windows machine wouldn't work, he claimed, because everybody knew Windows took several minutes to boot. Bachus and Blackley shared a look. It was time to play the ace up their collective sleeve. "This was the period when Kevin thought of bleem!," Blackley recalls.

In the months leading up to the beauty contest, Bachus and Blackley understood the difference between words and images on a PowerPoint, and a tangible proof of concept. Both men had connections at companies from AMD and Intel to NVIDIA and 3dfx, and cobbled together a prototype of the Xbox. "The thing I know how to do that many people don't is make stuff," Blackley says. "There was this new category of tiny PCs called Legacy-free PCs. They had tiny motherboards. They were very pure computers. They didn't have all the backwards compatible stuff that other motherboards had. It seemed like I could hack them to do instant-boot and modify Windows so it behaved more like a console."

Blackley converted one of the Legacy-free PCs into a proof of concept. Rather than unveil a generic-looking PC, however, he gutted the machine for parts and transplanted them in custom-designed, 3D-printed case. Bachus went a step further by installing bleem!, an emulator that ran PlayStation discs on Windows PCs with crisper graphics and faster load times. "I'm very proud of that," Bachus says of the demo unit he and Blackley assembled. "What we showed was held together with spit and baling wire, and we were constantly panicked that something would go wrong."

Taking out a copy of Tomb Raider, a third-person action game starring intrepid explorer Lara Croft, Bachus placed the CD in the tray of their Frankenstein's monster, picked up a controller, and played. "You could put in a software CD into it that it had never seen," Bachus explains, "but within seconds it was installed and letting you play the game. If you were a Windows developer, there was nothing different. All you had to do was build a Windows title. And just to be a bunch of dicks, we showed a PlayStation emulation program and put in a copy of Tomb Raider and started playing it."

Everyone stared silently as Bachus played. The silence stretched out after Bachus put down the controller and powered down the prototype. Gates turned to the WebTV team. Looking sheepish, they loaded a PowerPoint presentation. The damage had been done.

"That just goes to show that you never bring a PowerPoint to a demo fight," Bachus says.

Their presentation, adhering to the writer's adage of "Show, don't tell," influenced Gates, although not quite the effect Blackley intended. The demo unit's blink-and-you'll-miss-it boot time floored Gates to distraction. "Look," he snapped, turning to his executives. "These guys implemented this feature so it works. Why the fuck can't we do this with Windows? I've been asking for this!"

"Weirdly, our success at doing that derailed our whole message," Blackley says. "I don't even remember what happened after that. I think we did the demos and then everybody wanted to get the fuck out of the room. God, that's frustrating. You have to remember that we were in a big company that was a software company. Starting out with instant-boot was the right thing to do for a game machine, but politically not the right thing to do. That's the difference. That's a whole education."

Blackley was half right. Once Gates refocused his attention on the demo unit, he expressed interest and grilled Bachus, Blackley, and others. The guys were ready, supplying an answer to every line of inquiry. As the interrogation wore on, Blackley's respect for Gates rose. Like Blackley and Bachus, Gates had zero tolerance for bullshit. When he asked a question, they didn't prevaricate or attempt to feed him marketing spin. If they had an answer, they gave it. If they didn't, they said as much.

"I felt a kind of kinship with the guy. I was never scared of being asked questions by him, and always thought it was very interesting to see what he thought about stuff," says Blackley.

As the meeting wrapped up, Gates turned to Ballmer. "This is what we should do," he said, gesturing to the Xbox. "Can you make a business out of this?" Ballmer said he would. What he meant, everyone knew, was that he would delegate.

"Steve came back to Rick Thompson and me and said, 'You guys tell me if there's a business here,'" Robbie Bach remembers. "Rick then went and did that work for the fall."