(FORTUNE Magazine) – On the eve of its initial public offering, Goldman Sachs certainly has Wall Street's attention. It's the last of the great private investment banks to go public, and its IPO is the most alluring so far this year. Goldman will sell 11% of the firm during the first week in May. The offering will be priced between $45 and $55 per share and could fetch more than $3.3 billion. That would put a value of $25 billion on the whole company, making it the largest financial services IPO ever. But what's really got Wall Street matchmakers abuzz is what happens next: What will Goldman do with all of that valuable currency? "Think DaimlerChrysler," says Leo D'Acierno, a partner in the financial practice at Booz Allen & Hamilton. "You want to put together powerful institutions that have a similar ethos of believing that they're the best in their business. That means putting together Goldman and Fidelity."

Okay, stop right there. Goldman is going to make some acquisitions, but that's not going to be one of them, say sources very close to the bank. They told FORTUNE that Goldman has considered buying Fidelity but is daunted by Fidelity's size and the price it would have to pay. More important, Goldman's strategy doesn't include buying a large retail fund group to expand its downscale mutual fund assets. (Right now it manages only $57 billion in funds, sold through third-party vendors like financial planners and brokerage firms.) That would also knock out Janus, which may soon be spun off from Kansas City Southern. Its $108 billion in assets are primarily in commercial mutual funds.

Instead the 130-year-old bank will focus on what executives think are the missing pieces: fortifying its asset management division and building a strong Internet presence. Asset management is particularly important, since it will heavily influence Wall Street's perceptions of the company's earnings stability. After its IPO, Goldman will be valued as the nation's third-largest investment bank, behind Morgan Stanley Dean Witter ($63 billion) and Merrill Lynch ($33 billion). However, it's likely to trade at a price/earnings multiple below those banks' as well as other peers' like J.P. Morgan. Each of those firms has a large fee-generating asset management business that provides steady income streams when banking deals evaporate. The result is a more stable stock, and a higher P/E ratio. Goldman's earnings are too dependent on investment banking and volatile trading activities. (Asset management accounts for less than 10% of its revenues, compared with 17% at Morgan Stanley and 22% at J.P. Morgan.) If interest rates suddenly rise or an emerging country's economy, like Brazil's, tanks, Goldman's earnings, and its stock price, could fall faster and harder than its peers'.

Expect Goldman to be pretty picky about what it buys. Sources close to the bank say that the proper candidate will be a small to midsized firm with a niche investment strategy and a large institutional and wealthy-individual client base. A potential target would also need to fit into Goldman's spit-shined-shoe culture. Goldman will have to pay a premium for a company like that--perhaps 30% to 40% above current value. But it may be worth it. The firm wants to bolster the $277 billion in assets that it now manages for pension funds, foundations, and high-net-worth individuals. That is an important part of the bank's grand plan to become the premier "wealth manager" for affluent individuals. In fact, in the next couple of years Goldman will most likely carve out a practice within the firm to do just that.

T. Rowe Price Associates of Baltimore, always rumored to be a takeover target, would be a near-perfect fit. It manages $148 billion in assets--45% of which are for large institutions and elite investors. And, with a market cap of $3.9 billion, Goldman could easily afford to grab it. Amvescap, the result of the 1997 merger between Houston-based Aim and U.K.-based Invesco, would also fit well with Goldman's needs. It would be a lower-profile catch for the bank, but it has more assets under management than T. Rowe Price--$275 billion--and would give access to international investors. Another frequently mentioned target, though even more obscure, is Capital Research & Management. The Los Angeles firm runs the nation's third-largest mutual fund complex, the American funds. Goldman wouldn't even need to buy the whole asset-management company to get the clients it wants. It could follow J.P. Morgan's lead--that bank bought 45% of American Century Investments of Kansas City in 1997. "[Goldman] may buy a minority stake in a firm and skim off the top 401(k) and wealthy-client accounts," says Ben Phillips, a consultant at the Boston-based research firm Cerulli Associates.

An insurance play is another possibility. Marsh & McLennan is said to have rebuffed several would-be buyers of its Putnam Investments management group. But Putnam isn't the only big draw for Goldman Sachs. A steady stream of income from insurance fees would quell Wall Street's concerns that Goldman's sales are linked too closely to trading. "Marsh & McLennan has a hammerlock on the insurance brokerage business globally, and its asset-management group, Putnam, is clearly of the necessary stature," says Donald Putnam (no relation to the fund group) of Putnam Lovell de Guardiola & Thornton, an investment banking firm in San Francisco.

Any of these acquisitions would bring along the kind of retail mutual funds that Goldman doesn't want. A narrower but cleaner approach would be to buy a small bank or trust company like U.S. Trust or Chicago's Northern Trust. Their institutional and high-end individual accounts would fit in more easily with the Goldman gestalt.

To better access and service these tony clients, Goldman Sachs will also need a straightforward Internet strategy--something it admits it's lacking. The company took a small step in this direction in March when it bought a 22% stake in Wit Capital, an online banking and brokerage firm that offers its accounts access to IPOs that have traditionally been sold directly through brokers. But that isn't enough to make Goldman an e-powerhouse.

The bank may instead shop for an established electronic-trading group like Ameritrade or E*Trade, which is on track to have a million customer accounts by the end of this year. While this would open up Goldman to a lower-brow audience, the sacrifice would allow the bank to efficiently distribute securities and investment products to the wealthy clients it would eventually pick up with an asset-management purchase. Another possibility is Toronto-Dominion Bank. True, it is the fifth-largest of Canada's five big banks, but it also owns Waterhouse Securities, the second-largest online discount broker.

Some Wall Streeters wonder whether we will see the mother of all deals--a merger with Chase or Morgan Stanley Dean Witter. Others scoff. "Goldman's going to buy us? How's it going to do that?" asks a Morgan Stanley Dean Witter banker. He's got a point: Morgan Stanley's market cap is more than twice that of Goldman's. But given the waves of consolidation in the banking industry, Goldman needs to move swiftly if it wants to be a course-setter. Goldman itself would be attractive to at least several other financial giants. So if it doesn't shop quickly, it could wind up having to adapt to someone else's plan, not its own.