(FORTUNE Magazine) – John Doerr and William Randolph Hearst III could barely contain their enthusiasm. It was a Saturday morning in Palo Alto early last year when Doerr, arguably Silicon Valley's top venture capitalist, and Hearst, the scion of the Hearst media empire, sat down to breakfast with a NASA computer scientist named Milo Medin to tell him about a new company that was going to change everything.

Their new venture, to be called @Home, would deliver the Internet at blazingly fast speeds through the wires that now carry cable TV into 63 million American homes. Their efforts would transform the World Wide Web into a commercial mass medium, revive the sagging fortunes of the cable industry, and not incidentally, create a $2-billion- or $3-billion-a-year company that would make plenty of people--including Medin--rich. Already, @Home had the backing of the nation's largest cable operator, Tele-Communications Inc., and its CEO, John Malone. This would be another Sun Microsystems, another Compaq Computer, another Netscape Communications--all companies Doerr had helped launch.

Medin munched on his wild-rice pancakes and listened patiently. The 32-year-old engineer didn't know much about business--he'd ignored Doerr's first phone messages because he'd thought Kleiner Perkins Caufield & Byers, the venture capital outfit where Doerr and Hearst are partners, was a law firm--but he was reputed to be the best networking engineer around, so Doerr and Hearst wanted to hire him. Medin was, in fact, a legend among Internet pioneers, known for his brilliance, bluntness, and a stubborn streak. The story is told that he once shut down Internet access to all of Scandinavia after a Finnish hacker infiltrated a NASA computer.

That morning he was nearly as tough on Doerr and Hearst. "Nice idea," he told them, "but it's not going to work." The problem, Medin explained, was that building modems to hook computers to the Internet via cable systems wouldn't by itself result in superfast access. The sluggish performance that frustrates users of the World Wide Web stems from a host of factors--not just slow modems and standard phone lines but backlogs and bottlenecks throughout the Internet. What's more, Medin said, the Web's performance will only get worse as more users send more data through increasingly clogged pipes.

Right about then, Medin recalls, Doerr and Hearst looked as if he "had just run over their puppy with a truck." But he had good news too--there was a way after all for Doerr and Hearst to bring speed to the Web using cable modems. All they had to do was to build their own parallel Internet: a privately owned, high-speed network that would deliver to customers everything on today's Internet and more. He noodled the problem around awhile before figuring that @Home would have to spend between $120 million and $140 million to build such a network capable of serving ten million customers. "They didn't blink at that," said Medin, who then signed on to do the job.

Nearly two years later, @Home has grown, albeit haltingly, from a rough-hewn idea into a company in Mountain View, California, with 180 employees, its own coast-to-coast network, and yes, even a few paying customers--several hundred pioneers in Fremont, California, who, unlike the rest of us, are actually surfing, not crawling, the Web.

Whether or not the people behind @Home ultimately realize their grander ambitions--and a thicket of obstacles stand in the way--you've got to say this about them: They think big. Really big. Maybe that's only natural, given Doerr's stellar track record as a financier and Hearst's family history of empire building. "This is both a big idea and a big exercise in execution," says Hearst, 46, the grandson of newspaper magnate William Randolph Hearst. "I like to analogize @Home to Franklin Roosevelt's wanting 5,000 ships off the coast of Calais in two years." Says Tom Jermoluk, the former president of Silicon Graphics who took over from Hearst as CEO of @Home last summer: "I'm trying to build a service that will become ubiquitous."

Ubiquity remains years away, as we'll see. But any serious Web user has to be intrigued by @Home's promise to make jumping from page to page on the Web as easy as channel hopping on a TV. In theory, PC owners can download data at up to ten million bits per second, more than 100 times the speed of a phone line tied to a 28.8K modem and far faster than high-speed ISDN connections offered by phone companies. (The telcos are experimenting with better technologies, but these are still in the trial stage.) In practice, cable-modem speeds are slower but still fast enough to delight people like Mark Lieberman, a Sunnyvale, California, dentist. "I couldn't see living without it," he says. "I'm on the Web 100% more now because it's instantaneous." What's more, the cable modem is always "on," eliminating delays caused by the dial-up process and busy signals.

The cost? Thirty-five dollars a month, plus a $150 installation charge. That's not way out of line for Internet access, given that many users pay $20 a month for unlimited access and order a second phone line. It does, however, make @Home the priciest channel the cable industry has ever offered. The average cable subscriber now spends that same $35 a month for a package of 30 to 50 basic and pay channels. Subscriber fees for @Home are split between the cable operator, which keeps about $20 a month, and @Home, which gets the rest. If all goes according to plan, @Home will also control the initial screen subscribers see when they switch to the Web, so that the company can sell advertising and perhaps even take a slice of all those commercial Internet transactions due any day now. Yes, Doerr and Hearst and Jermoluk are thinking big.

They'd better be thinking small too, because grand schemes like @Home can be tripped up if little things, like installing the service and answering help lines, aren't done right. Bigger worries also loom: Debt-burdened cable operators must invest billions of dollars to upgrade their infrastructure for @Home to work. And TCI, which owns a 45% stake, recently reported disappointing third-quarter results and told suppliers to slow down shipments of new equipment to many of its systems.

Still, cable desperately needs the business boost that Internet access may provide. That's why Malone was receptive when Doerr, who dreamed up the concept (and the name) of @Home, presented the idea to TCI. Recalls Bruce Ravenel, CEO of TCI Internet Services: "Doerr brought his particular vision to the table that said, Silicon Valley startup, independent management, headed toward a public offering, and creating a nationally branded service that the cable industry would deliver." Doerr figured he'd get the whole cable industry behind @home, but that hasn't happened--major operators Time Warner (the parent of fortune's publisher) and Continental this fall launched their own Internet-access products, Road Runner and Highway One. But Comcast and Cox Communications have joined TCI as equity partners in @Home, each buying a 14% stake, giving the company the backing of three of the nation's five largest cable operators. Together TCI, Comcast, and Cox serve about 21.4 million cable subscribers, and they are marketing @Home to other cable operators too.

Lately, Wall Street has punished all cable stocks, none more than that of TCI, whose Class A shares have lost over 40% of their value in 1996, falling to lows not seen in nearly five years. The problem, analysts say, is that debt-burdened cable operators are under attack from direct-broadcast satellites (DBSs) and from telephone companies, at the very time when they need to invest in new services to compete. Worse, cable operators have lost credibility by failing to keep promises to deliver interactive television, digital TV with hundreds of channels plus telephony. "They have lurched from one delayed introduction to the next over the last few years," says Sharon Armbrust, a senior analyst at Paul Kagan Associates, a Carmel, California, consulting firm. "This is the last, best hope for the cable industry to set itself apart from its competitors. They need good, hard results." Industry leaders agree, although naturally they give this an upbeat spin. "This is a product I'm really pumped up about," says William Schleyer, president of Continental Cablevision. "It's a whole new market for us. It's a market we can own."

While Wall Street generally remains skeptical, some analysts who have taken a close look at cable modems are bullish. Richard Bilotti of Morgan Stanley places a $1.8 billion valuation on @Home--not bad for a company that so far has gone through about $50 million in startup capital. "This is a good product, and I think the industry is going to make it work," says Bilotti. Deployment of modems, he figures, should add 10% to 15% to cable systems' value by leveraging their big investments in upgraded, two-way plants. One caveat: he thinks @Home's rollout could take four to six years.

For its part, @Home isn't making predictions--perhaps because the company has already failed to live up to its own early hype. In the press release announcing @Home, Hearst promised to roll out its service by "early 1996 with broader deployment to follow shortly thereafter." Later came predictions that the company would sign up one million users in its first year. Hearst's resignation hurt the company's image too, though he'd called himself an interim CEO all along. A math whiz at Harvard, an early advocate of PCs, a magazine editor, a cable TV executive, and an investor in high-tech companies, Hearst had initially caused a splash by quitting his job as publisher of the Hearst Co.-owned San Francisco Examiner (tired) to join Kleiner Perkins and @Home (wired). Critics say he lacked focus, despite his brilliant theoretical mind; Jermoluk, by contrast, is an operations guy--just what @Home needs, as the company moves from theory to practice.

The primary challenge for @Home is now to get its product to market. TCI's Ravenel says, "The technology works. The demand is there. The service is great and customers love it. We're at the point where the name of the game is execute, execute, execute." He makes a good case: millions of frustrated Netizens are begging for bandwidth, and TCI, Time Warner, and Continental all report brisk demand for modems in test markets. Says Doerr: "I get letters from people who want to know where we're going to deploy next, so they can move there. No kidding." Should the Internet somehow fail to develop into a mass medium, all bets are off, but that's an unthinkable prospect to the folks at @Home. "Most people think that the Internet has been overhyped," Doerr says. "I think it's been underhyped. I think it's going to be more powerful even than television because it's two-way."

This year's technical and market tests have erased most doubts about cable-modem technology. Inside the home, the cable wire splits, with one line linking to the TV and the other connecting to an Ethernet card inside the PC. Data travels to and from the cable system's head-end, which in turn is linked to @Home's networks. Popular Websites are automatically replicated and stored on local and regional servers, so users don't have to wait for data from all the way across the Internet. The idea, Medin says, is not only to speed access but also to capitalize on the fact that storing bits has become a lot cheaper than transporting them. "It's classic computer science: Push the data closer to the user," he says. While cable operators who aren't partnered with @Home question the need for such an elaborate network architecture, Medin says its value will prove itself as Internet performance deteriorates.

The Internet-access market that @Home hopes to transform is rife with competitors; profit margins are already notoriously thin. The @Home team must contend with online industry leaders AOL and Netcom; Microsoft, with its vast resources and desktop dominance; the marketing pros at AT&T, Sprint, and MCI; and the Baby Bells, with their customer bases and reputations for reliability. Pacific Bell, for example, has signed up 58,000 California customers since launching an Internet-access service in May. Richard Hronicek, president of Pacific Bell Internet Services, says that's largely because reliability, not speed, is what matters most to users. "The Web is very exciting, but still the No. 1 use of the Internet is e-mail. That doesn't require speed," he says.

Ravenel disagrees: "So they've got 50,000 dial-up customers. That's good for us. Because dial-up, compared to this service, they're just night and day. When AT&T announces WorldNet [its Internet-access service], that's great news because it grows the base of people who go online and are frustrated with the experience they have today. Go for it, guys! Build market share! You spend your marketing dollars, AT&T, on convincing the masses that they should sign up for the World Wide Web. Then we'll switch them because we've got a better mousetrap."

Can cable ever deliver its mousetrap? An inherent problem with @Home is that responsibility for rolling out the service lies not with Jermoluk & Co. in Mountain View but with cable operators in places like Hartford and Baltimore. This creates an installation nightmare with all the appeal of Cable Guy II. Customers who want @Home must pick up the phone, call the cable company, and say, in effect, "Sure, send a guy over to rewire my house, pry open the back of my computer, and hook me up." That's no simple task, given the variety of hardware and software configurations on home PCs. Installation now requires two people--a cable tech and a computer tech--and can take hours.

Delivering customer service could prove even more vexing, particularly for an industry whose public image rivals that of the IRS. In fairness, cable is more subscriber-friendly these days--a little competition can work wonders--but customer perception lags reality. "Don't underestimate the public's hatred of the cable companies," says an industry analyst. Internet access offers cable operators a chance to redeem themselves but at a steep price. Seven-day-a-week, 24-hour-a-day, high-quality customer service isn't cheap.

Rivals don't think cable companies are up to the task. Says Don Hutchinson, senior vice president at Netcom: "Customer support will prove to be a defining issue as this industry moves forward, and I think it's been underestimated." For now, @Home will share the job with its partners, with the cable operators handling customers' easy questions and bumping up harder ones to @Home. "It's a challenge, but I think it's doable," says Tom Baxter, president of Comcast, which is revamping its entire customer-service operation. "A year ago, we would've answered phone calls in 65 places. Six months from now, we'll only answer them in eight places. And we've decided we need smarter, better-educated people." Morgan Stanley analyst Bilotti points to another worry; he calls customer service "the most volatile operating expense" facing the operators and warns that spending too much on it could "severely lengthen" the time needed for the cable-modem business to make money.

Scale, of course, could help resolve many operational issues. "If I don't get to millions of homes," Jermoluk says, "I can't make the thing pay for itself." Eventually, @Home wants retailers to sell "@Home ready" PCs with a built-in modem and cable-ready plug; besides simplifying installation, this would shift to consumers the cost of the modem, now around $500. Similarly, customer-service procedures created for the early rollouts can be replicated elsewhere. "You design once, and you implement cookie-cutter fashion," says Ravenel.

Fair enough, but @Home's reluctance to discuss rollout schedules raises questions about how long it will need to achieve such efficiencies. So do the troubles at TCI. Ravenel insists cable modems, unlike other equipment, will continue to be delivered as scheduled--but TCI is revising its plans to upgrade systems to the two-way fiber-and-coax architecture needed to optimize high-speed access. The upshot is that unlike DBS technology, which became universally available instantly, cable modems won't be an overnight sensation. By year-end, TCI plans to offer @Home in Hartford and suburban Chicago, Cox will bring it to Phoenix and Orange County, California, and Comcast will roll out in Baltimore. Beyond that lies uncertainty.

The slower the rollout, the longer it'll take to promote the @Home brand. More important, any delay will slow development of multimedia content that depends on a broadband pipe. Ultimately, @Home wants to deliver on this promise: to transform the Web experience by making it easy, rather than cumbersome, for users to enjoy rich graphics, TV-like images, and CD-quality audio. The screen that greets @Home customers when they turn to their Internet channel is a good example of what can be accomplished. The gateway is colorful and lively, with a constantly updated news feed and an interactive guide to the Net that even technophobes can manage. Says Jermoluk: "In the long run, don't tell anybody, but this is interactive TV.''

No wonder Web media companies, particularly those with video and audio assets, are rooting for @Home. "Cable modems are a huge opportunity for major providers to distance themselves from everybody else," says Halsey Minor, CEO of C/Net, which produces television and online content. Nevertheless, few programmers are developing content for @Home--understandably, since it's so uncertain how soon today's small broadband audience will expand significantly.

@Home has put together an impressive list of what it calls media partners, including USA Today, the Wall Street Journal, and HotWired. But what exactly defines a partner remains unclear, since none of the companies has agreed to a business model yet. @Home's strategy, Hearst says, is "to get enough mass to become an important distribution channel" and share in revenues generated by programmers. But Tom Hagopian, general manager of ESPNet SportsZone, a popular Website, says, "That proposition is a nonstarter for us." On the contrary--ESPNet would like a cut of the subscriber fees generated by @Home, just as ESPN's cable networks share in the monthly fees paid to cable operators.

So many worries, so little time. Never mind--the @Home people are undaunted. Doerr says @Home will become one of perhaps "five really great world-class companies, new companies, built around the Internet." Jermoluk says, "This is going to be like a rocket ship." It's possible they're right. If so, @Home will usher in a new era in cable--and not a moment too soon.