Perched near the pinnacle of a building in Rockefeller Center, Sirius Satellite Radio's luxe headquarters seem lofty enough to touch the sky. The sleek, hardwood-paneled space, with its glassed-in studios and an upper-level balcony for the C-suites, stands at the epicenter of New York's Media Gulch. It's a place fit for royalty, and this past November the satellite-radio company won itself a king, in the person of ex-Viacom president Mel Karmazin, whom Sirius crowned as its new chief executive.

For Karmazin, the coronation was quite an about-face. At Viacom, he had been one of satellite radio's foremost skeptics. But a falling-out with the lord of that media empire, Sumner Redstone, coupled with the lure of 30 million stock options and a $1.25 million annual salary, apparently eased Karmazin's doubts. Word of his ascension to Sirius's throne electrified the industry. Just one month earlier, Sirius had paid a king's ransom of $500 million over five years to secure the mouth of the self-proclaimed "king of all media" Howard Stern. Karmazin and Stern had built Viacom's Infinity Broadcasting into a radio powerhouse, and analysts hailed their alliance with Sirius as confirmation that satellite radio might actually hit the sweet spot on the public's dial. Weeks later, the testosterone was still flowing through Sirius's studios. "We're making news," crowed Patrick Reilly, the company's communications chief. "XM is buying ads."

XM would be XM Satellite Radio, Sirius's archrival, which is led by Hugh Panero, a battle-tested veteran of the cable-TV wars. From his headquarters in a gritty neighborhood in northeastern Washington, DC, Panero endured the headline-grabbing Stern signing, which curiously arrived during the very week that his own shock jocks, Opie and Anthony, began airing on XM. Then came the ballyhoo over Karmazin. "When we started out, all of the so-called experts said that no one was going to pay for satellite radio — and one of the loudest of those voices belonged to Mel Karmazin," Panero noted wryly. "We're happy to provide him with this new entertainment platform to participate in, as well as gainful employment."

Panero can afford to be a little cocky. For the past three years, XM has ruled the digital-radio airwaves. It is closing in on 3.1 million subscribers, each of whom spends $9.99 a month to sample upward of 130 music, talk, news, sports, and weather channels. Sirius, with a similar smorgasbord of offerings, lags with roughly 1 million subscribers, who pay $12.95 a month. XM was the first to market with a boom-box receiver (the SKYFi); it was first with a wearable radio (the MyFi); and it's selling far more subscriptions through its partnerships with automakers. Sirius, named for the Dog Star, has mostly endured a marathon season of dog days since XM leaped ahead of it.

The face-off between XM and Sirius might well be unprecedented in modern business history. In most industries, the competitive landscape is populated by many rivals. But satellite radio is truly a rarity — a duopoly of combatants slugging it out for potentially huge gains. Eight years ago, when the Federal Communications Commission awarded two licenses for satellite spectrum at a price tag of $90 million each to American Mobile Radio (now XM) and CD Radio (now Sirius), it created a perfect laboratory setting in which to observe raw, one-on-one competition.

The battle between XM and Sirius offers a classic case study in strategy and execution — where the sky is, literally, the limit. They're playing a kind of high-risk chess, with billions of dollars at stake. In a series of bet-the-company moves, each organization took a different line of attack in developing its technology, forming alliances, managing innovation, and making deals. While the game is still in play, it's already yielding some valuable lessons: that first-mover advantage is bunk; that unabashed risk taking must be leavened with pragmatism; that while strategy matters, everything depends on execution.

And there's more to this story. Even as both upstarts began to take each other on, they were simultaneously attempting to build the first new digital-media network since the Internet. At the same time, they had to outflank the oligarchs of the earthbound radio industry, such as Infinity and Clear Channel Communications, which had grown fat on broadcasting's consolidation. All in all, radio's space cowboys claim just under 4% of a nationwide audience of 100 million households. For XM and Sirius, the fight for air supremacy has barely begun.

Opening Gambit: Spend Furiously — but Wisely

Sirius is sometimes described as a scrappy underdog. It's more of an underperformer, generating a projected $68 million in revenue while piling up $450 million in losses in fiscal 2004. And it's a dog that lives like a fat cat.

Sirius has a 15-year, $4.9 million lease on its swank headquarters. The company binged on building showcase studios, which sat idle for nearly two years. Now workers are transforming a two-story glass enclosure — which dominates the lobby — into Stern's new digs. The company says that being in the media capital of the world helps it attract high-profile talent.

True, the real estate represents chump change in the scheme of things. The economics of building a digital-radio network dictate that Sirius and XM alike must run at a supersonic burn rate. It costs roughly $250 million to launch a single satellite (Sirius has three orbiting the Earth; XM has two). To store its countless digital-audio files, XM has assembled what's believed to be the nation's largest concentration of IBM servers. Like Sirius, XM is pumping red ink — in 2004 alone, it projected $360 million in losses on $220 million in revenue. Factor in other startup costs, marketing and promotion efforts, distribution deals, technology investments, and the price tag for hiring brand-name talent, and analysts estimate that both companies will spend well north of $2 billion before they ever make a buck.

Still, XM has done a better job of living within its means. Panero once ran a national pay-per-view TV network out of Denver — a media-industry backwater — where he learned firsthand that "in today's world, it's not about where you are, it's what you do." When it came time to build XM's broadcasting mecca, Panero bypassed the glitz of Manhattan for a neglected patch of Washington where the real estate was cheap, the talent market was booming, and Capitol Hill was just a short cab ride away.

XM is based in a century-old brick building that was once home to National Geographic's printing presses. When Panero first toured the sprawling, derelict warehouse, it was littered with broken glass and dead pigeons. He promptly signed a lease for $14 per square foot. (XM purchased the building in 2002 for $34 million.) "I didn't want us out in Dulles, in that graveyard of technology companies," he says. "I wanted us in a building with character that's near the vibrant city, so we could attract young people to work for the company."

First as an executive at Time Warner Cable and most recently as the chief of Request TV, Panero grew up in a business where the incumbents, at least in the early days, would rather kill off the competition than improve their own product. By the mid-1990s, he saw that the Goliaths of the radio industry were no different. Deregulation begat consolidation — today, Clear Channel alone owns or controls more than 1,200 stations — and radio had become a vast wasteland. The big radio companies homogenized the music, dumbed down their programming, piled on 20 minutes of commercials every hour, and pulled the plug on technology investments. Panero didn't need much convincing: Radio was ripe for revolution.

Seize the Center: "In-house" Means "In Control"

When Panero signed on in 1998, XM was the earthbound startup and Sirius was heading for the stars. Sirius was the first to be founded, it was first to launch its IPO, and it won the space race when it put its satellites into orbit almost a year ahead of XM.

But the business landscape is littered with first movers that were first to stumble, and Sirius is no exception. In its rush to make it to market, Sirius made a serious strategic error: It farmed out the development of the chip-set design for its radios to Agere Systems, a Lucent Technologies spin-off. An endless series of technical problems, coupled with the meltdown in the telecommunications industry, forced Sirius to cool its jets.

Panero followed a different strategy. Many times during his days in pay-per-view TV, he had seen how the cable companies would outsource their technology to a Pioneer or a General Electric, and then face interminable delays as the tech giants dealt with bigger priorities. He wasn't going to repeat that mistake. In what proved to be a game-changing move, Panero hired a crack team of Motorola engineers and kept XM's chip-development effort in-house. That allowed it to control costs, stay on deadline, and gain ground on a floundering Sirius.

(Sirius says the chip problems have been corrected and are now "ancient history.")

But Panero also understood when to look beyond his own walls. He decided to corral a heavyweight strategic partner that would lend credibility to the concept and amp distribution. He targeted General Motors. To Panero's way of thinking, GM was the trifecta: It was the world's largest carmaker; back then it owned DirecTV and Hughes Electronics (making it a satellite content and manufacturing company); and it had launched its OnStar navigation services. All of which meant that GM understood satellite radio.

But could XM, with its untested technology hosting an unproven business plan with unavailable funds, get GM? Several times, over the course of months, talks between the companies nearly cratered. Panero decided to bring it all to a head.

He gathered the various parties in a Maryland hotel — GM as well as DirecTV, Hughes, and executives from several large financial institutions, some 35 people in all — and set a midnight, do-or-die deadline. Through "one part creativity and one part perseverance," as XM chairman Gary Parsons put it, XM pushed each side toward the finish line. By the morning of June 9, 1999, they had a deal: GM and the rest of the consortium purchased a $250 million stake in XM. GM's buy-in created an instant halo effect for satellite radio, and Honda quickly allied itself with XM. For Panero, the GM deal was a lifesaver: "There wouldn't be a satellite-radio industry today if we hadn't pulled that off," he says.

Sirius employed a similar strategy — it too made a play for GM — but XM convinced GM that it would reinvent itself to meld seamlessly with the carmaker. It promised that all of its business processes — technology development, marketing, even billing and accounting — would mesh with GM's systems. "We couldn't expect the giant of the auto industry to change everything so it could incorporate our product," says Parsons. Sirius would eventually land partnerships with BMW, Ford, and Chrysler — and says it will rapidly expand distribution through automakers — but its technical problems initially forced it to rely more heavily on big-box retailers.

With its new receivers ready to roll in two GM models, XM dug out of the dotcom wreckage and was the first to hit the airwaves, in November 2001. "Everybody forgets that our competitor was better financed, they were a year ahead of us, and they were supposedly going to be first to market," grins Panero. "We shot by them because they didn't execute on their plan."

Sirius, stilled mired in its Agere problems, did take some action in 2001. It axed the company's founder, David Margolese, and brought in Joseph Clayton, the former president of star-crossed Global Crossing North America, to be its new chief. Sirius finally hit the air in July 2002, but the damage was done. With nearly a one-year head start, XM is leading the race to sign subscribers by a three-to-one margin. First with technology and then with distribution, Sirius stumbled badly. It has yet to recover.

Counterattack: Content Is King

Analysts predict that programming will be the next battleground in the satellite-radio war. Here, too, each company is blazing a somewhat different strategy. Early on, Sirius put the music front and center, and its eclectic lineup of channels has won it high praise.

Now it's trying to boost its stations' profiles by giving shows to anyone who has any kind of following, from Eminem to Basketball Hall of Famer and Grateful Deadhead Bill Walton. It's unclear whether Karmazin will change the mix. At Viacom, he told The Wall Street Journal that broadcasting's creative side amounted to little more than "arts and crafts."

At XM, arts and crafts — creating stars and must-hear channels — is the focal point of everything it does. Lee Abrams, who's widely credited (sometimes blamed) for pioneering the album-oriented format that dominated FM radio in the late 1970s and 1980s, presides over XM's music programming from an office bedecked with framed gold records. He's determined to put the magic — and the swagger — back into broadcasting's soundscape. He has given XM's DJs a game plan that flips modern radio's programming conceit on its ear: "We're not going to tell you what to play. You're going to tell us."

The result has brought an ear-opening array of music to the XM band, where a flick of the dial might take you from 1960s crooner Glenn Campbell to heavy-metal thrashers Lamb of God. But Abrams is after more than variety. He wants every station to have a distinct personality, from the Top 40 channel, which is as culturally significant as a shopping-mall food court; to the rap station, where if you're not part of hip-hop nation, you won't understand it; to the country-music station, which is ground zero for red-state values. Abrams calls this "point-of-view programming." By presenting people with a vast suite of stations, each with its own discrete sensibility, Abrams is betting that every first-time listener will find at least a few stations to relate to — and a reason to subscribe.

"Lee understands that radio does three things: It informs, it entertains, and it provides companionship," says Robert Unmacht of In3 Partners, a media-consulting firm.

In its music programming, he says, "Sirius believes that music is king, which is fine as far as it goes. But if you take a station that has personality and one that's all music, the personalities will win, because people want that companionship."

End Game: Diversify Your Deal Making

If there were any questions about Sirius's target demographic, the Howard Stern deal zapped them for good. Sirius is betting everything on the dudes — men between the ages of 18 and 49. The evidence: First, Sirius inked exclusive deals to broadcast the NBA and NHL.

It brought in the buxom one, Pamela Anderson, to do some promos. Then, when it had just a few hundred thousand subscribers, Sirius capped its sports mojo with a $220 million blockbuster to broadcast the NFL over the next seven years.

The Stern deal is in a league of its own. But was Sirius throwing a Hail Mary? While Stern's $500 million contract is a great win for him, the payoff for Sirius is harder to see. Though Sirius claims that Stern has a fan base of 12 million, the industry magazine Talkers puts that number at closer to 8.5 million. To offset his $100 million-a-year paycheck, Stern must convert 1 million of his listeners into Sirius subscribers.

Radio pros argue that Stern should not be underestimated. "The two biggest talents in talk radio are Rush Limbaugh and Howard Stern, and they are remarkably similar," says Unmacht. "If you strip away the act — the superconservative and the supersexed — they're doing classic radio. Stern will drag people across to Sirius."

XM, for its part, has pursued an entirely different theory of deal making, one that is alternately restrained and unfettered. For a company that has taken on a few of its own death-defying risks, XM can be remarkably pragmatic. Panero believes that the unsung part of deal making is having the discipline to make no deal at all. XM waited out Sirius's high-profile sports and celebrity agreements, and then it pounced. This past October, the very month that Sirius landed Stern, XM launched its own spending spree when it locked in its mammoth $650 million, 11-year broadcasting and marketing arrangement with Major League Baseball. Baseball will deliver many more hundreds of hours of programming than football, but it will also add something of equal value to XM's nonmusic channels: every demographic beyond the age of 12.

XM is working to ensure that its deals add up to a diverse lineup that will draw the widest possible audience. Shock jocks Opie and Anthony (infamous for broadcasting a live report of a couple having sex in New York's St. Patrick's Cathedral) are likely to draw a different audience than former NPR host Bob Edwards, after all. Sirius counters that it, too, aims for variety and says many of its channels appeal to women, kids, and older listeners. Add in deals for football, basketball, and hockey, and Karmazin will have the programming in place for a big boost in ad revenue from Sirius's nonmusic channels. Karmazin, after all, did as much as anyone to bloat conventional radio with commercials. He now has the chance to deploy the same off-putting strategy via satellite.

Panero is betting that the race for brand-name content will continue indefinitely. But as XM and Sirius drive to sign up an estimated 25 million subscribers over the next five years, it's likely that the showdown is entering a different phase — one where the outcome will depend more on execution than strategy. Now that each company has settled on its theories for technology development, partnering, programming, and deal making, they must take on the big question that confronts every entrepreneurial effort: Can you get it done? "It all comes down to the hard work of managing in a complex, multifaceted business," says Panero.

Media analysts generally agree that Sirius can prosper without winning. But author and stategy guru Gary Hamel disagrees, arguing that this is a zero-sum game. "These battles to build new markets look like a marathon," he says. "But they're more of a 100-meter hurdle. In the end, the winner is the one who clears all the critical hurdles that will turn the effort into a genuine business. Sirius stumbled at the very first one, and business history shows us that it's very difficult to catch up." Karmazin, who's generally credited as one of the toughest operators in broadcasting, could well find a way to kick-start Sirius and launch a whole new race. But right now, the sprint for the stars is XM's to lose.

Bill Breen is Fast Company's senior projects editor.