(MONEY Magazine) – Three years ago, Larry Ellison, CEO of database-software giant Oracle, took the stage during a gala press conference at Radio City Music Hall. Bathed in theater smoke and dramatic purple light as a laser show danced above his head, he told the crowd about a vision he had: The personal computer, he explained, was finished. And good riddance. PCs, after all, were expensive, hard to use, prone to crashes and difficult to service--and the latest model quickly became obsolete. A new and better kind of computer would soon arrive, he predicted, that would radically alter the computing landscape. Rather than being complex, self-contained units of processing power and storage capacity--dependent on Intel chips and Microsoft software--they would be like ordinary appliances: cheap, reliable, solid state. Yet they would deliver cutting-edge technology by tapping the remote computing brawn of a massive network. Oracle had just finished building such a machine, something Ellison called the network computer, or NC.

The NC, of course, was a spectacular failure, and the reasons for its downfall are, in retrospect, easy to recount. Ah, but what a difference three years makes. Because what was once a quixotic dream is now reality (though neither Larry Ellison nor Oracle have much to do with it). Later this year, computer makers including Gateway, Compaq and IBM will begin flooding the market with devices not far from Ellison's NC vision, putting us on the verge of a consumer computing revolution. The impact on technology industries--not to mention technology investors--could be huge.

Enabling all this isn't a new generation of Intel chips or Bill Gates' new Windows 2000. Rather it is the Internet--the network upon which these new machines will rely to deliver computing power, operating software and even data storage. Now known as Internet appliances, or IAs, these simple computers--some costing no more than a few hundred dollars--will allow even diehard technophobes to be e-mailing and Web surfing in just a few minutes. And that will be only the beginning.

It would be foolhardy to claim that the death of the PC is at hand, or even that it's likely--after all, previous generations of technology, like mainframes, remain significant long after new ones arrive. But we are undoubtedly entering a new age of computing dominated by the Internet, where having a PC is no longer a requirement. Welcome to the post-PC era.

THE NETWORK IS THE COMPUTER

The Net, in fact, has already altered the way we use PCs so significantly that it's hard to recall what an unconnected PC is good for. Not too long ago, most consumers used PCs as efficiency machines, to help make onerous tasks faster, from filling out tax forms to writing letters. (And, occasionally, playing a game of solitaire or Flight Simulator to blow off steam.) The Internet-enabled PC, however, encourages us to perform many of these conventional computing functions online. And it opens up far richer stores of information, entertainment, community, shopping and convenience.

The first generation of IAs that will be available this fall will attempt to capitalize on this Internet shift, catering to consumers who want Web access but don't want the cost, crashes and upkeep that come with PCs. Computer visionary Bill Joy, a co-founder of Sun Microsystems, points to the way wireless phones are used now as an analogy: A wireless phone without a connection to the wireless network is good only for storing numbers or jamming a door open. In essence, it's a dumb machine, and the majority of its intelligence is kept elsewhere, within the decentralized phone network. That, argues Joy, is increasingly true for computers: "The network is far more important than the hardware." And much like cellular phones, the cost of these machines may eventually be subsidized by network service providers like AOL or Earthlink in exchange for a service commitment of a year or two.

And this is all mere prelude to upheavals still to come. What happens when broadband Internet access (with connection speeds 20 times today's levels) becomes widespread, perhaps around 2003? At that point, there will be few things a PC could do that an IA couldn't. Plus: no more minutes-long dial-up rituals. IAs will always be on and as easy to use as the telephone. If you wanted to write a letter, for example, or crunch a spreadsheet (tasks beyond the range of the current Internet infrastructure), software companies called Application Service Providers would let you do that online for a small one-time or ongoing subscription fee. Your personal files will be backed up automatically and stored remotely.

Likewise, a greater proliferation of Web-enabled devices such as cell phones and personal digital assistants over the next few years will further remove the consumer experience of computing from the desktop. Indeed, market research firm IDC figures that Internet appliances, including mobile phones, interactive TVs, PDAs and video-game consoles, will outsell home PCs by 2002. (For a consumer update on the current nexus of cell phones and the Net, see "This Is Your Future Calling" on page 88. And for an investor's perspective on interactive TV, see page 53.)

800-POUND GORILLAS

So who has the most to lose in the post-PC era? The instinctive response may be the Wintel duopoly, since Microsoft's software and Intel's chips have dominated the definition of the PC--their gear is on an estimated 90% of the world's PCs. "The whole world is about to change," says Michael Murphy, author of California Technology Stock Letter, "and the No. 1 guy is usually the most threatened."

Already, hardware makers like Compaq, Gateway and IBM are partnering with firms like upstart system software manufacturer Be Inc. and chipmaker National Semiconductor to supply them with lower-cost products. And Internet giants like Yahoo! and AOL, whose sole purpose is to be always on your screen, are ready to step in with any software application you may need. For the first time since perhaps Netscape's IPO in 1995, people speak openly and confidently about a computing world where Microsoft is not dominant.

Indeed, the spiritual godfather of this new wave of computing may well be a 30-year-old Finn named Linus Torvalds, who earned his reputation largely because his creation has been the most successful attempt yet to subvert the long-established hegemony of Microsoft Windows software. A self-professed geek, Torvalds tweaked a few lines of a program called Unix 10 years ago, creating an operating system that his legion of followers acclaim for its fortitude (it goes months without crashing) and compatibility with a wide variety of hardware. Dubbing his work Linux, Torvalds posted the base lines of code on the Internet, free for all to use to create new applications. Since then, Linux has become a phenomenon, an increasingly popular operating system to run the corporate servers that make the Internet go. "I think the freedom and the flexibility that Linux gives is something a lot of companies are interested in," says Torvalds, who now works for a company called Transmeta, which manufactures energy-efficient chips for mobile Internet-ready devices. As early as last year, in fact, Microsoft CEO Steven Ballmer told money that Linux was the greatest threat his company faced, at least as significant as IBM and Apple in the 1980s and Netscape in the 1990s. Torvalds is often described as the anti-Gates because he not only gave his code away, he doesn't profit directly from its popularity. Instead, he simply marvels at what it's become. "Everywhere you turn, companies are starting to use Linux to some degree, because it fits well in the post-PC era," says Torvalds.

But counting Microsoft and Intel out would be premature. The fact that these are among the most profitable and highly valued companies in the world gives them leverage to move into--or buy into--key areas in this new revolution. That makes calculating their exposure to a post-PC world a tricky process. We've tried to lay out the risks and opportunities for them, as well as for a handful of other key tech titans--Apple, Dell, AOL and Yahoo!--in a series of sidebars that accompany this story.

The clear winners in this revolution are those infrastructure companies that stand to benefit from an increased reliance on the Internet: Cisco Systems, Ellison's Oracle, Sun Microsystems, EMC and Exodus Communications, to name a few. Unfortunately, all of these companies carry stock market valuations that are so high and so clearly reflect their dominance in the Internet arena that any material business gain from the rise of the Internet appliance is priced into their stock. If you don't already own them, the impact of IAs alone aren't enough to warrant buying such highly priced shares. (There are also some clear losers, which we highlight on page 84 in "Three Stocks You Should Avoid.")

So what should you consider buying to take advantage of this new wave of change? We've found five promising opportunities in four different sectors of the tech business, each of which stands to get a significant boost from the IA phenomenon. In the first three sectors--hardware, chipmaking and Web hosting--our picks are reasonably valued, relative to many other hot tech names. Our final two stocks, both component makers, have more extreme valuations but are small companies that we feel have a good chance to grow quickly enough to support their prices. Read on and see if you're convinced. (All data are as of March 27.)

HARDWARE: GATEWAY

Personal computers are not going away, since they will still dominate the business market for the forseeable future thanks to entrenched IT departments and corporate inertia. But given the attractiveness of IAs, it's hard to imagine that PCs will be the growth engine they once were.

Most of the major PC manufacturers have begun marketing Internet appliances. (One notable exception: Dell Computer; see page 82.) And that's despite the fact that selling IAs alone is not likely to be profitable. Compaq and IBM, for example, both lost money on their PC businesses last year; building IAs promises to be an even lower-margin endeavor. But the boxmakers do see potential profits in bundling their machines with monthly Internet service charges as well as advertising revenues from companies acting as portals. IBM, for example, has partnered with AT&T and Fidelity Investments: IBM will make the machines and configure them for AT&T's Internet service. Starting later this summer, Fidelity plans to give them out at low or no cost to customers who don't already use the Net. As we noted earlier, this is essentially following the business model in the mobile-phone business, where the hardware itself is a virtual giveaway, and the service plan makes up the cost and then some.

Makes sense. Unfortunately, we are not convinced that IBM is the stock to buy to take advantage of this "outside the box" strategy. Aside from the fact that consumer computers account for a small proportion of its business, making it a less leveraged opportunity, we think that IBM has a host of other complications (see "Warning Bells at IBM" on page 45).

Instead, we think Gateway is best positioned to deliver on this strategy. Gateway not only makes IA hardware, but it also delivers services through its own fast-growing ISP. Gateway began this year with about 400,000 paying subscribers, according to Credit Suisse First Boston analyst Mike Kwatinetz. By the end of the year, he expects that number to jump to 2.5 million. Two years ago, 100% of the company's profits came from selling PCs. Last year, the first under new CEO Jeffrey Weitzen, it earned 20% from non-PC related businesses, like ISPs. This year, the goal is to reach 30%.

Gateway's plan makes sense to AOL, which signed a deal valued at $620 million in cash and $180 million stock in order to get a high-profile spot on Gateway's PCs and Internet appliances. Gateway, meanwhile, will also hawk AOL subscriptions in its stores, sharing in the revenue from any new customers. On its retail side, the company plans to open another 80 Gateway Country Stores, giving it more than 300 locations to demonstrate just how these new and unfamiliar gadgets work.

Kwatinetz says that based on its "beyond the box" plan, Gateway, which recently traded at $56, could reach $78. It trades at 31 times the $1.83 per share that it is expected to make in 2000. Meanwhile, the consensus of analysts figure earnings to grow 22% annually over the next couple of years.

CHIPMAKERS: NATIONAL SEMICONDUCTOR

Intel is undoubtedly the cream of the crop when it comes to chip manufacturers. But in the new IA world, demand won't be simply for the most powerful chips. Instead, the chipmakers most likely to see near-term benefit from the IA movement will be those that have designed a suite of IA functions on one chip. That's why we like National Semiconductor: its Geode chip set, the size of a matchbook, provides Internet access and e-mail, in addition to video capabilities. Thanks to its functionality, size and quality, it's become the chip of choice to run many Internet appliances, including Compaq's and IBM's versions. In its third quarter last year, National Semi's orders for the Geode more than doubled, accounting for 3% of that quarter's $549 million in revenues. National Semi says Geode sales are expected to grow 100% over the next 12 months, making the company the leading player in this niche chip space. National Semi stock is trading at a recent $63, or 43 times next year's estimated earnings.

WEB HOSTERS: VERIO

As more and more software moves to the Net, outsourcers that run other large companies' websites stand to be among the biggest gainers. Forrester Research figures that the business of Internet access, website hosting and application hosting will quadruple in size to $88 billion by 2003. The market leader in this area is Exodus. But, as mentioned earlier, Exodus shares are anything but cheap: The company currently carries a stupendous $27.8 billion market cap on $242 million in sales. With so much new business rushing into the sector, we're confident that the No. 2 player, Verio, will also thrive--and its relative valuation makes it downright cheap. While Exodus is trading at a price-to-sales ratio of 118, Verio (with a market cap of $4.2 billion) is trading at just 16 times estimated sales.

Why the huge discount? Because Verio started out as an Internet service provider and because it serves the mid-size to small business market. And it will grow revenue 60%, on average, over the next two years, compared with Exodus' 150%. (Note that none of the companies in this sector post earnings yet; they are spending aggressively to build out the server farms that support their services.) Still, more than half of Verio's revenues now come from the higher margin Web-hosting businesses, and CFO Peter Fritzinger expects Web hosting to make up 65% of revenue by the end of the year.

COMPONENTS: SANDISK, SMARTDISK

These are the raciest picks of the bunch--smaller but faster-growing businesses with the opportunity for big upside moves as well as the peril of a painful collapse if they falter. We started by looking at memory makers that would benefit from the boom in Internet appliances. The big opportunity: makers of flash memory. In a Web-browsing device, flash memory takes the place of the disk drive, essentially letting you read, write and store data on a rugged, low-powered component the size of a stamp. And unlike other forms of memory, such as DRAM or SRAM, data aren't erased when the machine is turned off. The market leader in flash memory, with a third of the market's share, according to IDC: Sandisk. The company is expected to pull in $500 million in sales (twice 1999's $247 million) and earn 80[cents] a share this year--an 84% jump in earnings from 1999. The only problem: Sandisk is hardly a secret. Although it is expected to grow earnings at a 30% annual clip over the next few years, its shares currently trade for an astronomical 181 times expected earnings and 16 times sales. Clearly, not one for the faint of heart. Nonetheless, Jeff Wrona, manager of the PBHG Tech fund calls Sandisk one of his favorites.

The second component area we focused on: external storage. Until broadband connection speeds become the norm, most people who use Internet appliances will still want to store files like digital photos and word-processing documents locally. Our favorite stock: SmartDisk, particularly after it announced its acquisition of VST Technologies. VST makes drives that incorporate what's called FireWire technology, a high-speed connection that transfers files from your IA to an external storage device up to 30 times faster than USB, the current industry standard. The acquisition will double the sales of SmartDisk to $100 million and should boost earnings per share by a dime, to 50[cents]. Based on that estimate, its shares, which recently sold for $35, change hands for a lofty P/E of 88, or about three times the company's projected future growth rate of 30%--high but perhaps not too high a price to pay for a stake in the coming revolution.