(FORTUNE Magazine) – One sure place to find suits in ultracasual Silicon Valley is in downtown Palo Alto, on the sidewalk outside the offices of WebTV. Power guys in Hugo Boss arrive in Lincolns, their business cards marking them as heavies from, like, seriously grownup corporations. It's true that in April Microsoft bought the startup for $425 million, which can give a company a certain cachet. WebTV had also gained some notoriety last fall by launching a set-top box that brought the Internet's World Wide Web to TV. But the real reason the heavies are here is more important: WebTV has come up with a compelling glimpse of that Holy Grail in the technology and media business, interactive television.

In the past ten years, broadcasters, cable companies, Madison Avenue, and Silicon Valley have spent millions trying to take TV beyond the next pay-per-view Mike Tyson bout and $35-a-month cable bill. They desperately want to make over Americans' TV sets with the computing power, interactive programming, and connectivity that could turn them into gold mines for everyone from the local pizza shop to General Motors. Plugging the PC into a network sparked the creation of new billion-dollar businesses like 3Com and AOL; now the suits think it's time for TV to take the plunge.

The birth of this industry is shaping up as a spectacular clash. Cable TV companies are trying to make sense of their sudden suitors, the we-have-all-the-answers high-tech companies. The maneuvering is all about bandwidth (the ability to pump pixels into the home) and technical standards. Cable companies have the bandwidth--and will probably win big as upgraded cable systems make possible everything from TV Web browsing to video-on-demand to who knows what else.

The companies that win the standards battle (i.e., the ones that get to write the rules) will also make billions. At the center of the fight is Bill Gates, who believes Microsoft's future growth depends on capturing a big piece of the interactive-TV market. The math is simple: About 20 million households in the U.S. have PCs with modems, but more than 64 million have TVs connected to cable. It's going to be a hard run for Gates, though: He faces not just the usual Silicon Valley competition but also shrewd cable companies unintimidated by Microsoft.

Microsoft bought WebTV hoping the startup might have the magic touch. Its second-generation set-top boxes, which will be on sale this Christmas, link TV and the Web in a revolutionary way: You can watch a football game, see a small icon onscreen offering player statistics, click it with the remote, and view the stats. You can still see the game action, picture-in-picture; just click again to go back to watching the game on the full screen. Whereas last year's WebTV merely let users browse the Web, the new version integrates programming with Web pages. It's easy to see where Microsoft and WebTV can take this. Like this Disney movie? Click here to order Disney videos. Want to read that book Oprah is gushing over? Click to order from Amazon.com.

There is reason to doubt that interactive TV will blossom readily, of course. For one thing, it may take a long time to figure out what consumers really want from interactive TV--and how much they will pay. (WebTV has signed up only 150,000 users so far.) For another, we've seen would-be interactive TV revolutions fail before, such as Time Warner's costly Full Service Network trial in Orlando.

But this revolution is real--this time there's the Internet. "We never would have succeeded if we'd had to line up the content underlying WebTV ourselves," says WebTV co-founder Steve Perlman. That was the mistake Time Warner (the owner of FORTUNE's parent) made in 1994, when the Web was unknown. To provide pizza ordering on its Full Service Network, Time Warner paid a fortune for custom software that ran on expensive computer workstations attached to household TVs. Pizza Hut and consumers liked the service, but it was too costly to commercialize.

Today an ace high school Net buff can deliver the same software in HTML, the basic computer language of the Internet. More to the point, pizza ordering and countless other services have bloomed all over the Web; WebTV does not have to invent them. Better still, WebTV's new set-top box will sell for $299, vs. $349 for last year's version, reflecting the fast drop in prices and the rise in power of microprocessors. (WebTV service still costs $19.95 a month.) Competitors will soon make it possible to access the Web even more cheaply.

No one knows all this better than Gates. He wants to position Microsoft technology in every critical segment of the Information Highway--cable TV, delivering video over the Net, satellite broadcasting. That means insinuating Microsoft's Windows operating system wherever possible so that Microsoft will effectively be in control, just as it is in control of desktop computing. Future models of WebTV will be based on the Microsoft Windows CE operating system, a version of Windows adapted for use in consumer electronics. Gates also hopes to generate revenues by providing software and even acting as a middleman for online banking, bill paying, shopping, and other commerce on the Web.

What frustrates Gates is the slow pace at which the future is unfolding. In June, Microsoft startled Wall Street by paying $1 billion for 11.5% of Comcast, one of the largest cable companies in the U.S. The $29-billion-a-year cable industry had been in a funk for two years because of fights with regulators and competition from satellite TV, among other things. Gates' move signaled that he sees the cable industry as the best hope for kick-starting TV's digital revolution.

Why cable? Simply because the coaxial cable that runs past 95% of American homes is a pipeline without peer. It's incredibly fast, accommodates huge amounts of data, and is already there. In the past year or so, cable companies have begun switching to digital systems, which are even faster. In short, cable companies were starting to get it, and Gates saw that the moment was right. "We believe there is a great opportunity to sell more software as people become connected at higher speeds," says Craig Mundie, senior vice president of Microsoft's consumer platform division.

So far, Microsoft's investment has paid off in one important way. Wall Street's faith in the cable companies is somewhat restored, judging from the 21% increase in the Kagan Cable Company Index, which measures cable stock performance. But Gates has had less luck persuading cable operators to adopt an interactive system grounded in Microsoft software. In July he met with several large cable companies in New York City and pitched a plan to build set-top boxes that would contain Windows CE and connect to Windows NT servers at the cable company. Gates also suggested that Microsoft should get a share of the fees cable companies would charge for new interactive services the boxes would make possible. The cable companies, while intrigued by the vision of a host of new services, were less than ready to sign over the future to Microsoft. "Bill Gates would like to be the only technology supplier for this whole evolution," Tele-Communications Inc. CEO John Malone told shareholders at his company's annual meeting in August. "We would all be very foolish to allow that to happen. Bill has to accept the fact that he cannot have quite the dominance in supplying our industry that he has developed in supplying the PC industry."

The cable companies came up with a different idea: a process to develop open standards to govern the workings of the ultimate set-top box. A deadline for proposals was set for early October; CableLabs, the R&D center assigned by the cable companies to recommend a standard, received no fewer than 23 proposals from companies ranging from Intel to Sony.

Microsoft is going along with the process and has submitted its proposal, but that only makes CableLabs' job trickier. The problem is the unending rivalry between Bill Gates and Oracle CEO Larry Ellison. Oracle and its allies want set-top boxes using programming technology not controlled by Microsoft--namely, Sun Microsystems' Java and other "open" computer languages. Oracle owns a majority stake in Network Computer (NCI) of Redwood City, Cal., which has its own WebTV-like system. Last month RCA launched an NCI-based system on par with the first-generation WebTV; later this year Zenith will roll out a souped-up version competitive with the latest WebTV. "The only way to achieve simplicity and acceptable cost," says David Roux, an executive vice president at Oracle, "is not to have one supplier that provides everything but rather open competition based on pricing and innovation. The Soviet system didn't work in the Soviet Union; why would it work in the cable system?"

Microsoft, of course, sees things differently: It wants the industry standard to be Windows CE. The company claims CE works as a platform for many applications, as well as for several microprocessors. As a stable foundation that supports vital tasks like encryption and networking, insists Microsoft's Mundie, Windows CE will make it easier for manufacturers to build lots of functions into the projected set-top box. Microsoft will license Windows CE to all comers. But adopting it as a standard would put the evolution of the set-top box operating system entirely in Microsoft's hands. That would seem to defeat the purpose of creating an open standard through CableLabs.

Gates' worst fear is that CableLabs might recommend a unique operating system for the set-top box, a step recommended in several proposals. To add to his leverage, Gates is rumored to be negotiating another cable investment, perhaps in Tele-Communications Inc. (Given Malone's warnings about Gates, this is a show no one should miss.)

There's another question, perhaps the biggest: Can Microsoft and Silicon Valley learn to address TV viewers, that alien nation of passive beings who do not enjoy--let alone understand--reconfiguring their TCP/IP settings? It may be that WebTV and other Valley-centric offerings aren't what consumers want. Consider the product of Worldgate, a 60-person startup in Bensalem, Pa., that is another contender in the CableLabs standards race. The company is conducting trials with cable companies to provide Web access, as well as broadcasting linked to the Web, using a simpler, faster, and less expensive system than WebTV's.

Worldgate's service operates just the way cable always has, with a click of the remote and not much else. The system uses the latest set-top converter boxes cable companies are installing. Providing Web access to a few thousand households costs the cable operator only about $50,000, and the service could cost each cable subscriber as little as $4.95 a month. What viewers see on TV isn't as slick as WebTV's service, and Worldgate might never support add-ons like videophones. But that may not matter to the millions of households with no computer experience. After testing Worldgate in St. Louis, Charter Communications is happy and plans to start offering the service soon.

So maybe Microsoft and Silicon Valley aren't as smart as they think. To succeed, interactive TV will have to be cheap, compelling, and idiot-proof--no amount of techno-dazzle or evangelism will change that. "Microsoft had the idea that it was more sophisticated than the cable companies," says Cynthia Brumfield, an analyst at Paul Kagan Associates in Carmel, Cal., "but cable knows its business and sees no advantage in handing the business over to Microsoft." No doubt broadcasters--and Madison Avenue publishers and other industries in the ring--view Microsoft the same way. Some of those potential rivals are bound to be smarter than the likes of Apple Computer, and none is eager to leave money on the table for Bill Gates.