(FORTUNE Magazine) – Two years ago, Bank of America CEO Ken Lewis--who had portrayed himself as a cautious, tightfisted leader ready and willing to shun big, pricey acquisitions--shocked investors by unveiling a $47 billion deal for FleetBoston Financial. BofA's stock tanked as Wall Street roasted Lewis for paying a 40% premium for the troubled New England bank. But Lewis predicted that Fleet's heavy baggage of bad loans was poised to lighten, and boasted that BofA would cut yearly costs by $1.3 billion. He has delivered bigtime. Lewis vastly exceeded his target on expenses, and the former Fleet's profits followed his script. Since the merger was announced, BofA's stock price has risen 10%, waxing the returns of big rivals J.P. Morgan Chase and Citigroup.

Now Lewis is making what looks like another risky, overpriced deal by offering $35 billion for America's third-largest issuer of credit cards, MBNA. That's $8 billion (or 31%) more than MBNA's market cap the day before the announcement. In a single stroke, the merger will propel BofA past the current leaders in the card biz, Citigroup and J.P. Morgan, and make it the king of plastic with $142 billion in card receivables. On the surface, this deal--much like the Fleet merger--smacks of empire building. Sources close to the negotiations tell FORTUNE that MBNA contacted Lewis the day after it ended merger talks with Wachovia, BofA's fierce crosstown rival in Charlotte. Those negotiations collapsed when Wachovia declined to pay a premium for the card giant. Lewis then swooped in and agreed to terms within a week. "We concluded that paying a premium for a large credit card company is not consistent with our stated acquisition criteria," Wachovia noted drily after BofA triumphantly announced the acquisition.

What investors want to know is whether Lewis can score again by making MBNA pay for his shareholders. So far they're pretty much withholding judgment: BofA's stock dropped about 2% following the announcement. But a look at the numbers suggests that Lewis's plan is on the mark.

In addition to the $35 million purchase price, BofA (BAC, $45) will take a restructuring charge of $1.25 billion. So let's add that to the price, bringing the total to $36.3 billion. To gauge if that price is a ripoff or a bargain, we have to predict what MBNA will earn consistently once it's part of BofA. In 2004, MBNA posted $2.7 billion in net profits. This year the analysts polled by First Call predict that earnings will drop to around $2.5 billion. The main reason: MBNA took a big charge in the first quarter for the severance packages it granted employees in a recent round of layoffs. By 2007, First Call projects that profits will bounce back to $3 billion. Lewis claims he can pare $850 million in costs. If he succeeds, the former MBNA's earnings could reach $3.85 billion by 2007. Using that number, BofA is buying MBNA for less than ten times 2007 earnings--practically a steal. And BofA would be making almost $3.85 billion a year on a $36 billion investment, for an 11% return. "The nice thing about the deal is that BofA barely has to grow MBNA's earnings at all to make it a success," says Steve O'Byrne of Shareholder Value Advisors, a firm that specializes in calculating market valuation and consulting on executive compensation.

Still, Lewis faces two big challenges. First, he must achieve the promised $850 million in cost savings. His success in paring costs after the Fleet acquisition bodes well here. Second, he must help MBNA recover from some recent missteps. The company fumbled in 2004 by severely cutting back on zero-interest teaser offers for new customers, and it antagonized many loyal customers by doubling or tripling their rates and imposing big fees when payments arrived even a few days late. It infuriated others by switching them to American Express cards without permission.

MBNA has already reintroduced the teaser rates, and it's being far more careful about sticking it to good customers for late payments. "MBNA cost itself a lot of growth," says BofA's CFO, Marc Oken. "This deal will be a success if MBNA's revised strategy works, and we think it will." If BofA can boost sales of MBNA cards through its coast-to-coast network of 6,000 branches, the deal will rise from a modest success to a home run. Considering Lewis's batting average on acquisitions, it probably wouldn't pay to bet against him.