Last year, the magic word among Internet stocks was definitely e-commerce. But in 1999, say analysts and others, the magic word is infrastructure, focusing not on companies that are selling on the Web, but on the companies that help the so-called e-tailers to set up shop. The prospects look good for the coming year, except for the nagging suspicion that Internet-stock valuations could crash to earth this year.

For Steve Sigmond, analyst at

Dain Rauscher Wessels

, the top infrastructure pick is

Verisign

(VRSN) - Get Report

, a company in the digital-certificate business, which allows people and companies to verify one another's identities. "You'll see just an increasing adoption of new digital-certificate technology by both businesses and consumers," Sigmond says. Sigmond has a strong buy on the company and a price target of 100 -- about 25 points above where it is trading now. (Dain Rauscher was an underwriter on Verisign's IPO.)

Verisign's digital-certificate technology will be better integrated into Internet software in 1999, more widely available and easier to use, Sigmond says. "This is the first year that we'll see signs that this is mass-market stuff."

David Cummings, managing director of the

Noble Financial Group

, a New York-based investment bank and private equity firm, says the market's focus will shift from companies that sell on the Web to companies that can provide e-commerce companies with infrastructure and support. In the IPO market, he says, this trend will show up in brick-and-mortar companies designed to fulfill orders for online retailers, such as

On-Demand Solutions

, a subsidiary of

CMGI

(CMGI)

. Another example is

WebTrends

, which just filed to go public. This company's software helps people track usage of their Web site. E-commerce support is "sort of the next wave in Internet stocks," Cummings says.

Among Internet stocks,

Credit Suisse First Boston

analyst Bill Burnham sees a new category of company emerging between the online commerce companies and the technology companies that develop tools for e-commerce. The activity in this middle layer, he says, is outsourcing -- packaging technology in a form that companies can use without having to immerse themselves in the technology. Companies in this area are

Digital River

(DRIV) - Get Report

, which outsources software sales, and

Pilot Network Services

(PILT)

, which handles Net security.

"They're young companies," Burnham says. "The world is their oyster. As long as they can continue to execute, they've got tremendous opportunities." CS First Boston has an underwriting relationship with Pilot, which Burnham rates a strong buy; the brokerage hasn't been an underwriter for Digital River, on which Burnham does not have a rating.

But for anyone investing in Internet stocks, the risks loom large. "With the stocks I cover," Sigmond says, "the primary risk that they face in the development of a business model is whether these potentially huge markets ever actually develop." He adds, "I'm feeling very positive about overall market development, especially in the infrastructure space."

Instead, the risk in 1999 is valuation-related: for example, how the market values companies whose path to profitability is lengthy and not well marked.

"There's a big what I call dead cat underneath the rug that people trip over and pretend is not there, which is the lack of business models on the Internet," says Mark Stahlman, head of New York media research and financial services firm

New Media Associates.

"

It is likely to come to people's attention when they realize what the venture capitalists have already figured out," he says.

And it's going to happen in 1999, Stahlman says, "because the venture capitalists figured it out in 1998."

Citing the same gap in perceptions between private markets and public money, Cummings agrees. "No question about it, a lot of venture firms and private equity firms are no longer willing to fund 'me too' companies," he says. "Meanwhile, retail investors are chasing after those very same companies in a game of stock-of-the-week."

What will happen to public markets when investors change their minds? "It would be foolish for me to predict what a market based on fantasy will do," says Stahlman. Likely, it's a mixture that's enormously volatile. "Stocks that show a promise of early profitability will be absurdly strong," he says. "The handful of anchor investments probably will continue to be strong. And a lot of other people probably will be decimated."

"People are going to come back down to realistic valuations," Cummings says. In 1999? "Definitely."

David Braunschvig, managing director of

Lazard Freres

, declines to predict a downturn in the Internet sector. But given its high volatility, says Braunschvig, if there's a downturn in the broader stock markets for any reason, "there'll be an even sharper downturn in the Internet sector."

But for the moment, demand far outstrips supply. Before the late-December run-up in Internet stocks, Braunschvig estimated that a universe of 70 "pure-play" Internet companies had an aggregate market cap of $140 billion, but the available pool of stocks for trading, or non-insider holdings, was about $40 billion. "It creates an undersupply of stock," he says. "And that, in and of itself, is a factor that explains a lot of the increases." As more companies come on the market and the tradable market cap increases, that should ease up the undersupply, he says.

Burnham agrees: "What you've got here is a fundamental oversupply of demand." Investment bankers are going to issue supply, he says, in the form of IPOs. He estimates it will take about a year and a half before supply catches up with demand.

Eventually, Internet mania will turn ugly, says Jim Balderston, Internet industry analyst at

Zona Research

. "When it stops, it's going to stop dramatically," he says. "When the Internet fairy dust gets knocked off some of these stocks, the correction will be more dramatic than the rise in these stocks."

Probably. In the past, as we know, reports of the imminent death of Internet mania have been premature. "These stocks have continued to rise despite naysayers like me," Balderston says. "It's entirely possible it will continue in the new year."