REDMOND, Wash. (CNN/Money) - Talk about a success problem.

Microsoft (MSFT: down $1.93 to $44.30, Research, Estimates), despite its ailing stock price, ended its fiscal year in June with $38 billion in cold hard cash, plus another $14 billion in equity holdings. That's more than the value of all of AOL Time Warner, which, with a market cap of $47 billion, is the 28th largest company in the S&P 500 (and parent company of this Web site).

This is a perennial problem for Microsoft and one its top managers are certain to be grilled about Thursday at the company's annual analyst day at its headquarters. Even with the billions Microsoft has washed down the drain in failed investments (think: AT&T (T: down $0.06 to $9.00, Research, Estimates)), its cash figure went up by $6 billion in the past year.

The problem is what to do with the cash (an issue MONEY Magazine tackled in a spring feature -- click here). Some investors would love Microsoft to return the money to them in the form of a dividend. But as Michael Mauboussin, an analyst with Credit Suisse First Boston, argued in an excellent report this week, two shareholders who'd hate to see a dividend are Bill Gates and Steve Ballmer. Because dividends are taxed as ordinary income, Gates and Ballmer would have to pay taxes at a rate of 38.6 percent, compared with 20 percent for long-term capital gains.

In Mauboussin's plan, Microsoft -- and other companies with huge cash positions -- should announce regular, planned, predictable and generous buybacks. This would act like a dividend because it would be returning money to shareholders by removing shares from the market and improving the value of each share.

And it would be done in a more tax-efficient manner for everyone. Mauboussin estimates that Microsoft easily could soak up $28 billion worth of stock without hurting its long-term cash needs.

"As an investor, I want to have my cake and eat it too," says Mauboussin, who has been documenting the general demise of dividends in corporate America. "If you can achieve the same objective in a financially sound way, why not do it?"

His colleague, software analyst George Gilbert, predicts a different -- and intriguing -- use for Microsoft's cash. Gilbert thinks Microsoft will start aggressively providing financing to its customers, a technique industrial companies like Deere and General Electric have deployed for years.

This strategy has its downsides, as telecommunications equipment makers Cisco Systems and Nortel Networks found out the hard way. But imagine in these credit-constrained times if Microsoft would step up to the plate and declared itself a bank with $50 billion-plus in assets. That'd be a shot in the arm for an ailing technology economy.

Watch for highlights from the analyst day in my Friday column.

Remember Applied Micro Circuits (AMCC: down $0.29 to $4.09, Research, Estimates), which I dissed here in November, when the stock sold for about $12, or 22 times forecasted sales? Shares in the San Diego-based maker of telecom semiconductors trade for a little over $4 today, and Tuesday Applied Micro said it would cut its workforce by 25 percent, or 275 people.

RECENTLY BY ADAM LASHINSKY       Avoiding the value trap AOL's cookie jar The 'real' options problem

When a company slices and dices its headcount by a few percentage points, you know it's repositioning or trying to dial down to the right level. When it lops off a quarter of the company, announces plans to shut a chip-making plant and slows development of an important next-generation product, you know management hasn't a clue where the business is going.

To review, in November Applied Micro traded for 22 times Wall Street's revenue target of $165 million for the year ending in March 2003. But the stock was already down 85 percent from its highs, so somebody out there thought it was cheap. Now, with a market value of about $1.2 billion, it's trading for about 10 times its reduced expected sales of about $125 million. Subtract its $1 billion in cash and Applied Micro is worth somewhere between one and two times its expected sales.

Is it a value now? Who knows? Merrill Lynch analyst Mark Lipacis doesn't see the company earning profits at least through fiscal 2004. Can you put a value on reportedly estimable intellectual property and a $100 million-plus stream of profitless revenues? If so, then buy this stock.

Adam Lashinsky is a senior writer for Fortune magazine. Send e-mail to Adam at adam_lashinsky@timeinc.com.

Sign up to receive The Bottom Line by e-mail.