(FORTUNE Magazine) – So you have an idea for an Internet startup. It's a big idea. Well, not earth-shatteringly big, but of sufficient mass to justify, say, a blowout billion-dollar IPO. You put up a Website to prove your concept has legs. And as it attracts more and more visitors, it also starts to draw the attention of the megaportals, like America Online, Yahoo, and even Microsoft. As your site's ramping up, you look around and see a lot of other startups tinkering with the same idea. You get a little edgy. Six months later Microsoft sends people over to talk about a possible licensing deal, and before the discussion is over, a $400 million offer is on the table. They want your company. All of it. The offer knocks you back a bit: In your heart you don't want to sell, but you know that if you don't, one of those other startups will. And that IPO is still at least a year off, if you even make it that far. What do you do?

This dilemma is becoming more and more common in Silicon Valley, as the established portals indulge an insatiable hunger for acquisitions, inhaling startup after startup and assimilating their customers and technology (like the Borg in Star Trek absorbing new species into their own). As this colonization of the Internet becomes more complete--as the portals lay claim to more of the Net's prime real estate--the question facing startups is an urgent one: Are you strong enough and smart enough and useful enough to stake out your own piece of turf, however small, or do you need someone to take care of you?

Paul Deninger, CEO of Broadview International, a technology investment bank, gives the following advice to entrepreneurs: "If you have the opportunity to be the Cisco of your market, you are crazy not to go public. But if you will not be the Cisco in your space, you should sell, because your ability to generate long-term shareholder returns is limited. History shows that the cards are stacked against you."

Indeed, startups that do try to go it alone as a public company are looking at some nasty odds these days; markets are simply moving too fast, and outfits that won't sell often find themselves left behind. (For more on the speed of the new M&A, see the previous story.) True, 1999 looks like a record year for high-tech IPOs, with 279 deals through the first three quarters, but an astounding 2,637 technology acquisitions were completed during the same period. Even if you take just the tech acquisitions worth more than $100 million--the really serious ones--they, too, outnumber the IPOs (see chart). And it's by no means clear that companies that do go public are better off. According to a recent study by Broadview, nearly half the tech stocks that have gone public since 1992 are currently trading below their IPO price. Miss your numbers for a quarter and you may see 90% of your market capitalization--and your fortune--go down the tubes.

What you decide to do when the portals come calling depends on who you are, what your company has to offer the market, and how much you're willing to gamble. The trifecta, of course, is to start a company, go public, and then be acquired (which is how GeoCities got $4.5 billion from Yahoo). As Dave Roux, a partner at high-tech LBO firm Silver Lake, points out, for companies looking to be acquired "the best possible marketing is to take the company public. Done correctly, it can be a giant FOR SALE sign." Giles McNamee, an investment banker at Hambrecht & Quist (and brother of superinvestor Roger), agrees: "If you have the option of going public, then someone who wants to take you out has to compete against that pricing."

Few can pull that off. Instead, startups are increasingly being used as bait, testing the waters to gauge the interest an idea or business plan generates among the public, potential investors, and potential acquirers. That naturally gives rise to a broad spectrum of motives--apparent and hidden. At one end of the spectrum are the two-bit entrepreneurs just trying to keep an idea alive long enough to make a few million dollars off it. At the other are those who will settle for nothing less than building the next Amazon.com. And then there is the gray, middle zone, where the vast majority lie.

None of this is news to the giant portals. It's a Darwinian arrangement that suits them fine. To a certain extent, industry leaders such as America Online, Cisco Systems, and Microsoft are just outsourcing part of their research and development to Silicon Valley, letting the free market apportion R&D budgets to different products and independent teams. The market also provides a mechanism (called customers) to validate the results. Then the AOLs and the Ciscos use their stock as currency to buy the startups whose products are most in demand; those products then contribute (theoretically, at least) to future revenues and profit growth, giving the behemoths more currency to buy more startups. Many of the companies that aren't bought simply shrivel up and die. In no industry is this state of affairs better established than in data communications (for more on datacom, see the following story on Cisco's acquisition machine), but it is fast becoming the case in other Internet sectors as well.

The trick is to be able to make yourself irresistible. You've got to have something identifiable to sell. Generally, a portal buys a startup for one of two things: Either it is buying the eyeballs the startup has already managed to attract (which translate into advertising dollars), or it is buying a "sticky" new technology that it hopes will make the eyeballs roving its Website stay longer, come more often, and maybe even go shopping.

To stay competitive, though, the portals need to offer Web surfers every cool new feature that comes along. That faddishness tends to create a wave phenomenon whereby companies offering certain Web features are bought in quick succession. For instance, in August 1998, Amazon.com bought Junglee for its software, which can search the Net for the best prices on any particular item; within the span of a year, other private startups with similar comparison-shopping software were bought by AOL (PersonaLogic), Microsoft (CompareNet), Infoseek (Quando), Inktomi (C2B Technologies), and CNet (KillerApp). Other categories in which the portals have been picking up startups during the past two years include companies focusing on e-mail, chat, instant messaging, Web-based calendars, personal home pages, and e-commerce Website services.

Given the money at stake, it's no surprise that more than a few entrepreneurs in Silicon Valley are looking at the market as a game whose point is not to build a company but to create a little value--for themselves. In other words, more and more tech companies are being built to be bought. Roux calls them "designer startups," adding dryly that they are often "features masquerading as products masquerading as companies. You typically design one of them to be a 24-month gig--then you flip it and drip it" (like a slaughtered lamb). There is even a San Francisco consulting firm, Akwirus.com, dedicated to helping such startups get snapped up.

These are, in the end, marriages of convenience. The Microsofts, AOLs, and Yahoos of the world--desperate to find an edge--get what they need in the form of the techno-flavor of the month. The startup crowd gets what it needs in the form of cash and (nicely liquid) stock options. Sure, some of them are more principled than that. They want to be their own boss; they want to kick Bill's ass. But in the face of all that cash, eventually all but the most idealistic (or stupid) entrepreneurs will capitulate. "What you find," explains venture capitalist Steve Jurvetson of Draper Fisher Jurvetson, "is that when entrepreneurs are faced with their first major windfall, it is a test of their mettle. Do they want to make money or change the world?"

Not so long ago, practically any private startup that opted to be acquired instead of going public was considered a failure. Real companies went public--it was as simple as that. But to a large extent, the reverse is true today. It was Hotmail that opened the floodgates for Internet acquisitions when it agreed in 1997 to be acquired by Microsoft for $400 million in stock. That was an almost unheard-of price back then for a cash-bleeding startup with hardly any revenues. But e-mail is the most popular application on the Web. What Microsoft bought was nine million users of Hotmail's free e-mail service. Today 45 million people use Hotmail, and it delivers more traffic to the rest of the Microsoft Network than any other source.

Microsoft's shares have tripled in value since the acquisition, outstripping the returns of most IPOs. Still, in retrospect, Hotmail could probably have commanded a heftier market value if it had waited to go public, given that its subscribers were growing faster than virtually any other Web company's. At first, founder Sabeer Bhatia resisted acquisition offers by others, such as Yahoo and Excite. "Bhatia had an almost hallucinogenic optimism about his company," remembers Jurvetson, one of his VCs. Jurvetson used to joke with Bhatia, "If you don't sell, in a few years you will be able to buy Microsoft." But then Yahoo snapped up rival Four11, and Bhatia could see the writing on the wall. "It was clear Microsoft would have bought one of our other competitors or built it on their own," he says. Rather than tangle with the beast, Bhatia became one of the first people on the Internet to figure out how to turn nonpaying Web surfers into real money: He sold them to someone else.

Since that deal, the landscape has changed. An acquisition is becoming a more viable alternative to an IPO. In part that is simply because, thanks to the booming stock market, the portals' stocks have become highly coveted. But to a large degree, the taboo has started to lift for the startups--at least privately. Publicly, of course, hardly anyone will admit he'd like to be bought. It's like admitting you're hunting for a rich octogenarian to take care of you. Everyone claims to want to be a "stand-alone" entity, but the protestations are often no more than a charade. The first rule of the acquisition game is to pretend you're not playing. Many companies are starting to play, however, and how they go about it depends on who they are.

First there are those who will not consent to be acquired at any price. Take Kana Communications, one of the best-performing IPOs of the past couple of months--up nearly 400% since September. Kana makes software used by corporations to manage their torrents of customer e-mail. Started three years ago by Mark Gainey and his dog, Kana, the company's goal is no less than to bring customer service to the Web. At first Gainey went door to door to big corporations, asking them for their old customer e-mail so that he could study it. He put the e-mails in a database and started experimenting with ways to generate automatic responses. As the idea caught on, Gainey was approached several times with merger and acquisition offers, but, he says, "To merge did not feel right. It did not feel like we would have fully capitalized on the opportunity." Now he has a suite of software products and about 150 customers, ranging from eToys and Priceline.com to Chase Manhattan and Gap.

Next to Kana on the spectrum of acquisition candidates are companies with a realistic chance of pulling off an IPO but that for good reasons decide not to try. Hotmail is a prime example; another that contrasts more starkly with Kana is Webline Communications. Both Kana and Webline make software that helps businesses communicate with their customers. (Kana's software automatically responds to customers' e-mail queries; Webline's lets service reps in call centers share the browsers of customers phoning in for help surfing a company's Website.) Both get about $100,000 per client for licensing their software. Both were partly funded by Draper Fisher Jurvetson. Both even briefly considered merging with each other. But on Sept. 22, Kana went public with a $1.4 billion market cap, while Webline was acquired by Cisco for $325 million in stock.

Webline could easily have been in Kana's shoes. When Cisco came along, it had already picked Goldman Sachs as its lead banker for an IPO (Goldman was also Kana's banker). And if Kana's lofty market cap (now up to $1.9 billion) seems to validate its decision, Webline's CEO, Dan Keshian, asks an important question: "How many companies go public, have a great valuation, then miss a quarter, and all that value goes someplace else?" Kana CEO Michael McCloskey zings back, "There was no way that Webline was a sustainable company. It was a tool. Eventually, those companies don't survive."

Sometimes, laying the formal groundwork for an IPO (and getting the backing of a high-visibility lead banker) is enough to start the bidding, as was the case with Webline. Sometimes, though, it just takes a company with a good idea--a leader in its niche with good prospects for going public a year or two out, once its business model is refined. Web-calendar company When.com was at that stage when AOL snagged it for stock worth more than $200 million last April. Founder Ted Barnett explains, "We designed When.com with either outcome in mind: IPO or acquisition. Companies that tend to create the most value are the ones that are going for the IPO, because then you have a credible alternative when you are negotiating with your acquirer."

When.com was appealing to AOL not so much because of the eyeballs that came with it but because it was one of those extra-sticky applications. Barnett had taken a desktop application that people look at every day and moved it to the Web. Why would anyone want to keep a personal calendar on the Web? Well, it's free, it can be accessed anyplace in the world that has an Internet connection, and it can be shared by groups. And, unlike a desktop calendar, it can be linked to databases that track movie releases, IPOs, concert dates, sporting matches, trade shows, and TV listings. When.com has the potential not only to keep AOL customers glued to its site longer but also to help pinpoint advertising. (Going to a trade show? Here's a deal on an airline ticket.)

Barnett knew that the raw dollar value of his company after an IPO would probably have been larger, but that was at least a year or more off and there were tremendous risks in getting there. He didn't know what the company would look like by then, or even if he'd be running it. Also, the market was getting crowded: Literally dozens of calendar companies were popping up. But like Hotmail's Bhatia before him, he feared a larger danger looming. "There are only so many big portals out there," he says. "After a while each one would have picked its solution." Already Yahoo had acquired WebCal, and Amazon had picked up PlanetAll. A mere three weeks after AOL bought When.com, Microsoft pounced on another calendar company, called Jump.

Finally, there are the "designer startups," such as Mirabilis, with its popular ICQ ("I seek you") instant-messaging Web software. The software allows anyone registered to create buddy lists of other registered users and see whether they are surfing the Web at any particular time. In June 1998, AOL paid nearly $300 million for the Israeli company, even though it had no sales and not much of a business model--other than to build the number of users as fast as possible. The service did not even run any ads before the company was sold. One of the founders, Yossi Vardi, once told Bhatia, "ICQ's sole goal was to be like Hotmail." By that measure, Vardi did a good job. Twelve million people were using the free ICQ software at the time of purchase. That's about $25 a head.

Today, as the Internet overlords fortify their fiefdoms, private startups are going after smaller scraps of real estate. And it is getting harder for them not to look like acquisition bait. These companies range from serious Kana-like contenders to coy designer startups. "For some companies," says Gainey, "the most important product they are creating is the company itself. They ask themselves, 'What is the next thing Yahoo will need?'"

Regardless of whether they're just trying to bring eyeballs to the screen or creating "the next e-mail," it is getting harder to imagine they'll make it alone. In the eyeballs category, you'll find the consumer-product-review sites (Epinions.com, Deja.com, Productopia), e-mail list dispensers (eGroups.com, Topica), roving chat services (Hypernix's Gooey.com), and community sites (eCircles.com, Bolt Media). The next e-mail, on the other hand, is probably to be found in adapting common desktop applications for the Web, in areas such as file storage (i-drive.com, DocSpace), bookmarks (HotLinks Network, Clickmarks.com), event planning (Evite, seeUthere.com), or a seamless integration of all of these (Visto, Desktop.com). Each of these companies has a cool Website that could fill in some gaps for one of the megaportals.

Even in this world of more or less undisguised avarice, one does occasionally stumble upon an idealist--or at least a partial idealist. Jeff Bonforte, CEO of i-drive.com, knows, for example, that as Web file-sharing sites such as his attract more users (he's already providing Web-based file storage for MP3.com, and he's working on a deal with dozens of universities), they will also draw the attention of the megaportals. He knows there will be a flurry of acquisitions. And he thinks that if i-drive.com can survive, it should be able to go public at a much higher valuation than if he sells out now. He knows, in short, that he stands to make some pretty serious scratch. But he also knows that's not the whole reason he's doing this: "It would have to be a really, really good offer before I decided to sit in a cubicle somewhere in Building 27," he says. His chief technical officer, Tim Craycroft, a Netscapee, agrees: "I saw Netscape buy company after company and watched the acquired technology disappear and the people vest in peace."

Probably many Silicon Valley hotshots would like to think they had the chips to turn down an offer or two. You might like to think so too. But as Broadview investment banker Javier Rojas puts it, "You can say no until the numbers become interesting." Most of a founder's wealth, after all, is locked up after an IPO, whereas he can pretty freely liquidate stock received from a buyout. And even if the founder wants to go for the IPO, there is always pressure to sell from the vice presidents and investors, not to mention the spouses. In the end, just about everyone has his price.