This image was lost some time after publication.

By all rights, Facebook CEO Mark Zuckerberg ought to be feeling drunk with power right now. He has, I'm told, term sheets in his hands from the three giants bidding for a small piece of his startup: Microsoft, Google, and Yahoo. All three, I understand, meet his demands for a staggeringly high valuation on the company — $10 billion or more. Piled up behind them are countless offers from venture capitalists and private-equity players who would be content merely to have their funds' names attached to the untouchably hot social network. So who will Zuckerberg choose?

Microsoft, whose CEO, Steve Ballmer, is rumored to have flown in and met with Zuckerberg on Friday, has an advantage: It has control of Facebook's U.S. banner-ad inventory through 2011. The deal is lucrative for Facebook in the short term, netting it about half of its $150 million in estimated revenues for this year. But few understand how costly the deal could be in the long term.

The vogue these days is to speak of "advertising networks." But there's some truth to the faddish term. Advertisers naturally gravitate to the players who can distribute their ads as widely as possible. And like other networks, online advertising enjoys network effects. The larger the volume of ads one serves, the more data one generates about which ads are effective. That, in turn, lets an advertising network target ads more effectively, and price them more profitably.

Facebook has developed what it believes is breakthrough ad-targeting technology that exploits the wealth of personal information its users share. But by mortgaging its ad inventory to Microsoft, it has lost its testbed for improving this technology, and missed out on these network effects.

Google and Yahoo are approaching Facebook with similar deals: They promise to buy up Facebook's international advertising inventory for years.

Why would Microsoft, Google, and Yahoo be so willing to throw money at Facebook? Not, as some have suggested, simply because they're desperate to get a foothold in social networking. These companies are not run by dummies. They want to learn from Facebook, it's true — but they can take their learnings and run the results across their entire ad networks, improving their targeting. And they'll also build up relationships with large advertisers. The long-term benefits will accrue to Facebook's new minority investor, in other words. And Facebook? Once the deals expire, it will have to start from scratch.

If Zuckerberg's smart, he'll scatter all three term sheets to the wind. Perhaps he can cut a deal with Microsoft, but one that restructures Microsoft's banner-ad deal, making it as best a backstop to Facebook's own advertising efforts.

Facebook.com — the social network itself — may well be a fad, as Microsoft's Ballmer has suggested. We can only imagine how fast he had to backpedal from that statement, but it's true.

But Facebook the company is more than a social network. If Facebook's technology is everything its engineers promise — and granted, that's a rarity, but supposing it is — then Facebook needs to escape the Microsoft trap and start selling its own ads, beyond just the sponsorships its small sales team focuses on, as soon as possible. That technology, and the data it generates, could well make Facebook a permanent player in online advertising, the expert in hypertargeted, personalized ads.

I suspect Microsoft, Google, and Yahoo's dealmakers understand this. The question is whether Zuckerberg does — and whether he has the nerve, the verve, the swagger and swerve to believe in his company's abilities.

Watch what he does. If he takes an investment bundled with a long-term ad deal, then he agrees with Ballmer that his company is a fad, and he just wants to sell at the top of the market. If he restructures the Microsoft deal, shortening its term and curtailing its scope, then we'll know Zuckerberg has drunk his own Kool-Aid — and found it intoxicatingly sweet.