Of course, it was former CEO Steve Ballmer who spearheaded those mergers, and failed to make them profitable. Now, his successor Satya Nadella is effectively trumpeting a new era of Microsoft (MSFT) as an expert acquirer and integrator by announcing the purchase of social media icon Linkedin (LNKD) for $26.2 billion in cash.

It’s the biggest deal in Microsoft’s history. It’s also one of the most expensive transactions in the annals of technology. What matters to shareholders isn’t the happy talk about how “the likelihood of seizing more of the TAM (total addressable market) will increase through differentiated experiences,” but the numbers. And those numbers show that Microsoft will need to either generate gigantic improvements in Linkedin’s profits, or steer many millions of Linkedin customers to its own products, to make the deal a financial success. The chances are slim.

To see why, let’s examine the returns Microsoft generates today, and what it must do to generate the same margins on all the new assets its piling on its balance sheet via the purchase of Linkedin. At the end of the first quarter, Linkedin had shareholders’ equity or net worth—the difference between what it paid for its assets and what it owes to lenders and suppliers—of $4.6 billion. Microsoft is paying $26.2 billion for that equity, and assuming another $2.6 billion for miscellaneous liabilities not included in that number. Hence, it’s piling a total of $28.8 billion in new Linkedin assets on its books.

For this analysis, we’ll use a measure developed by Jack Ciesielski, author of The Analyst’s Accounting Observer, called “cash operating return on assets,” or COROA. It’s essentially the total cash a company generates each year before taxes and interest, divided by dollars invested in the software, working capital, customer lists, and all the other assets that produce that cash. Last year, Linkedin produced operating cash flows of $827 million on its then asset base of $6.9 billion (based on average assets for the year). That’s a decent return, or COROA, of 12%.

But for the new acquirer, the bar just rose enormously. Because it’s paying a super-rich price, Microsoft’s returns on LinkedIn, once it becomes a unit of the software giant, will be measured not on $6.9 billion in assets, but on what Microsoft is paying for all those new assets, that $28.8 billion. LinkedIn is actually running its business less profitably than the far larger Microsoft. Linkedin’s COROA is 12%, while Microsoft’s is 15.2%, meaning it generates just over $0.15 in cash on each dollar of assets.

So for Microsoft to lift Linkedin’s performance to match its own, Linkedin’s basic business, or that business plus the improvement it brings to Microsoft, would need to generate operating cash flows of $4.4 billion, or 15.2% of $28.8 billion. The problem: That’s more than five times the cash that Linkedin is making right now.

Given the size of these hurdles, it’s unsurprising that the slide deck that accompanied the announcement is long on generalities extolling the power of social media, and short on disclosing what Microsoft is adding in new capital, and what it takes to make that load of new capital pay for its shareholders.

That’s not to say the deal is another Nokia, or necessarily destined for failure. Microsoft may tame what looks like undisciplined spending at Linkedin, including huge equity awards, and could use Linkedin’s customer list to boost sales of its own services.

As usual, we don’t know what Microsoft will do, but we do know what Microsoft needs to do, if shareholders are to benefit. The deal is a gigantic windfall for Linkedin’s investors, who watched their shares jump 50% on June 13. The best bet is that their gain will be Microsoft investors’ loss.