Joost's office on the south end of Manhattan is in a geriatric 11-story building served by elevators so slow you suspect someone is pulling them up hand over hand. There's no company sign on the door, and inside, it's all urban hip, with exposed two-by-fours, open ductwork, and walls covered in silver metallic latticework that looks as if it were stolen off a futuristic chaise longue. I slip through subwaylike turnstiles at the front desk—yes, it does seem odd—and bend myself past a copier hidden by the kind of heavy red curtain you might see on a puppet-show stage. A public-relations woman then leads me into a squished conference room furnished with cheap plastic Ikea chairs. I know because I own the same chairs.

Joost has only about 30 people in Manhattan, plus 20 in London and about 60 in the Netherlands. Two years ago, when the public first heard about the company—at the time under its code name, the Venice Project—the startup was expected to use the internet to reinvent television, or render it obsolete. Joost had money, clout, and enough mystery to electrify the media. CBS joined an investment group—along with venture firm Sequoia Capital and Hong Kong billionaire Li Ka-shing—that pumped $45 million into the company. Joost was supposed to be really big by now. A YouTube. An iTunes. But it's not.

Into the Ikea-outfitted conference room walks the man who's trying to change that. Mike Volpi, 41, took the Joost C.E.O. job a year ago. Before that, he was the best-known executive at Cisco not named John Chambers. Volpi is an Italian who spent much of his childhood in Japan and most of his adult years in Silicon Valley. He is tall and angular, wears modish black-rimmed glasses, and maintains an appealing scruffiness—not that perfect Justin Timberlake stubble that seems to have its own greenskeeper, but more that of a guy who works on a loading dock. Despite the rough-hewn look, Volpi comes across as unusually mischievous for a C.E.O., especially when his infectious smile spreads across his face in the manner of a Wallace and Gromit Claymation character.

"When I came here, we didn't have a human-resources policy," he says, sounding more amused than alarmed. "We didn't know how many people worked for us. We had no real income statement or balance sheet. The first thing I had to do was put those things in place." But while he was doing that, something happened to Joost that no one expected back in the Venice Project days: It became an also-ran. "This market changed so much in a year," Volpi says. "It went from 'No one thought it was feasible' to 'Everyone is doing it.' "

What's "everyone" doing? Delivering free, ad-supported, professionally produced video over the internet in a way that protects copyrights, pays people for their work, and serves up something akin to traditional TV advertising. In other words, network television on the Web—C.S.I. via WiFi. The content, delivered on demand, looks great—not like some pixelated YouTube clip, but as good as or better than cable television. When Joost started, nobody could do that. Today, a bucketful of companies, ranging from ABC and Microsoft to Blinkx and Brightcove, can—and a newcomer, Hulu, can do it really, really well.

Something else unexpected happened to Joost. Or rather, didn't happen. While more than 10 million people are swarming to YouTube's user-posted videos every day, consumers are not yet falling in love with high-end internet television. Only 10 percent of internet users watch full-length shows on the Web. Fewer than 6 million people have downloaded the software that allows Joost to run on a computer—which gives Volpi a total potential audience smaller than that of Animal Planet's Puppy Bowl IV.

Still, nearly everyone in television and the tech industry believes that internet TV will be a multibillion-dollar business. Market research firm iSupply says advertisers will spend $3.3 billion on it by 2010, and the major networks are placing their bets. CBS has its money on Joost. NBC and Fox have pledged money and content to back Hulu. ABC is attempting to turn ABC.com into a destination video site. Even former Disney C.E.O. Michael Eisner is on the bandwidth bandwagon, investing in Veoh. Network executives all know that internet TV will reshape their business, even if they don't yet know how—or how the sites will make a profit, since none are even close to doing that now.

Hoping to goose both Joost and the demand for internet TV, Volpi is making two major moves. This year, viewers will be able to watch Joost videos in a browser window. Go to Joost's website, click on shows like Seth Green's edgy Robot Chicken or an old Rocky and Bullwinkle episode and you can watch them as easily as you'd watch a video on YouTube. Previously, all Joost users had to download and install special software. "The download may seem like a small barrier," says Brad Hunstable, co-founder of user-generated video site Ustream, "but it's a huge mental barrier."

Volpi's other gambit is more daring. Beginning with the N.C.A.A. basketball tournament this spring, Joost plans to carry live programming. That's a technological nightmare, but Joost's software may make it the first company that can send live shows to millions of users. Imagine Florida retirees from Chicago watching Cubs games live over Joost. Or soccer-crazed kids in U.S. suburbs tuning in to European soccer matches.

"Live TV will be a milestone for us," Volpi says.

Here's how fast Joost went from superhero to life support.

June 2006: Word of the Venice Project leaks out. It gets immediate cred because Niklas Zennström and Janus Friis, the guys who started Skype and the music-stealing site Kazaa, are behind it. All anyone outside the project's inner circle knows is that it will offer internet TV. Given Zennström and Friis' disruptive track record, industry watchers predict the death of network television.

July 2006: The Venice Project gets its $45 million in funding. The mystery surrounding the company fuels blog chatter worldwide. It's a tech version of speculating about the finale of Lost.

January 2007: The name Joost is revealed. Americans pronounce it juiced. The Scandinavians say yost. A few thousand people try the service in private beta. Their verdict: It looks cool, but there's almost nothing worth watching on it.

June 2007: Volpi, who was on Skype's board, is hired as C.E.O. At tech-industry cocktail parties, people debate whether he's the next Eric Schmidt.

October 2007: Joost is available to anyone, but new high-quality video sites begin popping up like prairie dogs in the desert. Hulu is unveiled; it raises $100 million from private equity firm Providence Equity Partners. ABC serves up videos on ABC.com.

December 2007: Adobe releases Flash Player 9 Update 3, which makes it easier and cheaper to display high-quality, professional video in a browser. Hulu is all over it; Joost is not.

January 2008: Influential tech website NewTeeVee posts a blog entry titled "Five Ways to Save Joost."

We're in an Italian restaurant in Lower Manhattan, a few blocks from Joost's office. The waiter brings over a chalkboard listing the specials. "Let me fix your Italian," Volpi says, erasing a letter with his finger. He can speak Italian and Japanese; his English has a slight lilt.

After ordering, we make small talk for maybe a half-hour before I bring up Joost's disappointing performance thus far. "Some thought we would instantly be transformational," Volpi says. "Those who know how media work had other expectations. I think it will be transformational—just not in six months."

He compares internet TV today to cell phones 15 years ago. People then didn't realize that mobile phones weren't replacements for regular phones; they were a whole new way to talk. Likewise, Volpi says, Joost isn't meant to replace broadcast TV. It's not even meant to substitute for TiVo, or what Volpi calls catchup TV. Instead, Joost's goal is to deliver video that people otherwise would have no way of seeing, whether it's vintage episodes of Mannix, jazz videos, Turkish dancing, Asian table-tennis championships, or new programming like Robot Chicken. "Just doing TV on the internet won't work," he says. "We're aggregating lots of niches." Some of Joost's hottest shows right now are old Transformers cartoons. "If we do it right, it will be an extraordinarily large business."

Joost's way of delivering video is unique. The system is based on the peer-to-peer technology that Zennström and Friis developed for Skype. When you watch a show on Joost, it is being reassembled from millions of pieces that are coming from other Joost users scattered everywhere. In milliseconds, your computer puts the pieces in the right order and displays the video in the Joost player. Simultaneously, it forwards the pieces to other Joost users who want to see the same video.

Joost's most significant deal is with CBS. Users can see a dozen or so CBS shows, including C.S.I.: Miami and clips from David Letterman's monologues. The network isn't cranking out any new programming for Joost; it's all repurposed fare. Joost also has shows from Comedy Central, and the N.B.A. serves up week-old games. Joost's most popular channels include Sexy Clips and Motors & Babes. The typical Joost user is a 25-year-old male. "They're watching sports, comedy, and anything with girls in bikinis," Volpi says.

The business model for Joost—and most of its competitors—is straightforward. Whenever someone watches a show, Joost runs a 30-second ad for every eight minutes of content and divvies up the money paid by those advertisers with the show's owner. Joost won't reveal the exact split.

Joost has always sold itself as a grownup version of YouTube. It has none of the user-uploaded content that can create copyright horrors or make an advertiser cringe when its product is placed next to home video of someone barfing. (Yes, you can find plenty of that on YouTube.) Joost's system protects content so that no one can steal it and ads can't be skipped. Also, videos can be blocked geographically, so if CBS has the rights to show Numb3rs only in the U.S., no one can use Joost to watch it in Poland. Joost is supposed to be a safe moneymaking haven for content owners and advertisers, yet neither top-quality programming nor ad dollars are cascading its way. As a private company, it does not report financial results, but as of early 2008, it had just 44 advertisers.

And now it has serious competition.

I'm at another restaurant, in Los Angeles. This one is called the Lemon Moon Café, and it offers a lot of exotic quiches and salads. Across from me is Jason Kilar, the C.E.O. of Hulu. He is smart and chipper and perfect-looking, like the president of the richest frat on campus. (The restaurant was his choice.) Kilar, 36, grew up at Amazon under the tutelage of C.E.O. Jeff Bezos. He worked there from 1997 to 2007, launching the company's DVD business and becoming a senior vice president. The most significant lesson he learned under Bezos, Kilar says, is "focusing on a world-class consumer experience." He has brought that to Hulu, which is a clean, simple site. You navigate by scrolling through thumbnail graphics of shows, which include 30 Rock, House, The Office, and The Simpsons. Click on a show and you see menus listing episodes. Click on an episode and it starts playing. It is much more like a mainstream TV experience than Joost.

Hulu had an advantage in starting up in mid-2007. By then, Adobe had begun work on a high-def version of Flash Player 9. The result: Hulu requires no download and no separate player. Thanks to Kilar, Hulu has another advantage over Joost. A show on Hulu can be embedded in any other website. The embedded Hulu video has commercials—served up by Hulu—that are tracked and counted and then paid for by advertisers. Consumers can watch Hulu videos on Hulu.com and on other sites that aggregate videos, such as Veoh and AOL. Bloggers can drop Hulu shows like King of the Hill into their blogs, and no matter where in the U.S. you watch a Hulu video on the Web, Hulu counts it and charges advertisers.

"The big challenge is having content find its audience," Kilar says. "It was important to me to do it virally. With Peter [Chernin, Fox TV's boss] and Jeff [Zucker, head of NBC], it was a quick conversation. They instantly got it."

Later, I ask Zucker why NBC and Fox invested in Hulu rather than going with Joost. "I was not that comfortable with Joost," Zucker says. "We wanted greater control over how our content was presented and distributed, and the key things advertisers and consumers are looking for are ease of use and a safe haven." By safe haven, he means a site where no one will stumble across those barfing videos or even Motors & Babes.

Hulu plans to be more than just a Fox and NBC outlet. "We're talking to ABC, CBS, and Viacom," Zucker says. "We're optimistic they'll be a part of it."

Even if those networks sign on with Hulu, that doesn't necessarily mean Joost has been squeezed out of the picture. No network is likely to sign an exclusive deal with a website—the networks want their shows on every viable outlet. Their strategy is the equivalent of ESPN's wanting to be on Comcast, DirecTV, Verizon FiOS, and any other system that delivers TV signals. No single internet-TV outlet is likely to be a total victor and command the field the way eBay dominates online auction selling. Still, when it comes to luring A-list content, Hulu has wrested the momentum from Joost. Kilar is too much of a pleaser to say it, but it's clear he thinks Hulu is ready to kick Joost's butt.

And what of internet video's million-pound gorilla, YouTube? "We presume you don't want to watch a 30-minute video on YouTube," says David Eun, who runs all of Google's content businesses. He's sitting in a top-floor suite at the Mandalay Bay Hotel in Las Vegas during the Consumer Electronics Show, where he's spending a lot of time talking to mobile companies about getting YouTube onto cell phones.

Nevertheless, Google will elbow into internet TV from the advertising side. No technology yet exists that allows a computer to understand what's in a video, then categorize the content for search and advertising purposes. In other words, the internet has no version for video of Google's powerful search technology and its AdSense advertising engine. Right now, if you search for video, the search engines read the text and keywords surrounding the video but not the video itself.

Google will change that. It has software currently used mostly to help content owners identify copyrighted videos on YouTube so that they can be removed. But, Eun says, that software can be adapted for video search and ad placement. He wants to build an "advertising machine for video" that rivals AdSense, the behemoth business that places ads beside text on Google and other websites. Content owners will be able to make money by having Google serve up ads wherever their videos appear. Consumers will find video by going to Google and typing in the name of the show they want to see. In February, Google announced the first tests of just such a system.

Today, the internet-TV space looks like a chaotic asteroid field. A couple of years ago, everyone expected Joost to dominate. Now you can't keep up with the new options. By early April, Blinkx expects to introduce a Joost-like offering. Microsoft is marketing its Silverlight video software to drive internet-TV sites, and this summer, Silverlight will power thousands of hours of video on NBC's Olympics site. For the next year or two, consumers may be baffled about what to find on which sites.

Eventually, however, internet TV will be important. Masses of consumers will watch shows on laptops. Computer chips in future TVs will pull video off a Joost or a Hulu and play it on big flat-screen sets in family rooms. Internet TV will change television as much as iTunes has changed music. But Volpi is right: It won't happen this year. Maybe not even in the next couple of years. And it's still an open question whether Joost can flourish—or at least become successful enough to get better furniture for its conference room.