As I write this, it's been almost a week since J.P. Morgan Chase & Co., with a degree of candor unusual for a public company, announced plans to terminate its $5 billion outsourcing contract with IBM Corp. and take back 4,000 employees previously transferred to IBM as part of the deal.

Since then, the business and IT press has been commenting busily on Morgan's decision, and the reactions range from claims that it is a "major blow" to IBM and the effort to move to utility computing, to hand-wringing over what it means for the future of "megadeal" IT outsourcing.

Yet certainly the real motive for killing the deal with IBM is the operational excellence of Bank One, and once Morgan completed the acquisition of Bank One, the writing was on the wall. From a business point-of-view, the suspension of the IBM contract is just one of the many uncertainties that accompany a highly charged M&A environment.

Under CEO Jamie Dimon and CIO Austin Adams, Bank One, itself the result of a long string of mergers, spent billions to beef up its information technology, and by all accounts, it was money well spent. So despite the disparity in size between the two—Morgan has upwards of three times the assets of Bank One—Bank One's IT looks to be the winner in the deal, and the reason the combined company decided to kill the IBM contract.

As for what it means for the technology side, John Parkinson, the chief technologist for the Americas at Capgemini and a columnist for this magazine, notes, "there are only a couple of hundred corporations big enough, and good enough, that outsourcing isn't attractive to them."

Size matters: In Parkinson's analysis, midsize companies will soon find it uneconomical, and uncompetitive, to try to do their own IT. But for the largest companies, the phrase "good enough" is critical. And clearly, having bought what Parkinson believes is "probably the best information technology executive team in the financial services industry," Morgan saw not just savings in taking its IT back from IBM, but plenty of competitive advantage as well.

Is it really a setback for the emerging concept of utility computing? "Well, it certainly demonstrates how difficult it is to do utility computing for real," Parkinson notes drily.