Computing Power Sold Like Electricity

Palmisano throws IBM's might behind a business for the future

Samuel J. Palmisano has seen the future of computing. On Oct. 30, at New York's American Museum of Natural History, IBM's CEO unveiled what his company deems The Next Big Thing: e-business on demand. Palmisano calls it the most important initiative IBM (IBM ) has undertaken since the Internet hit the scene. He's throwing all the might of IBM's 350,000 employees and $5 billion research and development war chest into the effort. "We have rallied the entire technology community at IBM around this," he says.

It's a call to arms much like the one delivered by Microsoft (MSFT ) Chairman William H. Gates III when he wrote the legendary 1995 "Tidal Wave" memo that galvanized the software giant to move beyond PC computing and on to the Internet. Palmisano is hoping IBM's "e-business on demand" will have just as profound an impact on the way corporations buy and use technology.

Palmisano's vision is a grand one. He sees a world where companies will no longer need to worry about technology. Computing power will be dished out just as a utility supplies water to customers who pay only for what they consume. Software will connect as easily as Lego blocks, and digital elves will spot potential problems and fix them before they happen. "This is exactly what customers want," Palmisano told BusinessWeek. "It's going to drive phenomenal productivity."

Behind Palmisano's new strategy are five key initiatives. He's banking on work already under way at IBM to push open standards such as the free Linux operating system and to push Web services. He's marketing so-called grid software for sharing computer resources. His troops are trying to convert customers to a pay-as-you-go utility model where IBM turns on the processing power and customers are billed based on how much they use. And in the research labs, an effort is under way to develop software that automatically fixes bugs. Indeed, the first seeds of IBM's labor are coming to fruition. In an April upgrade of its Tivoli network management software, IBM engineers built a feature that automatically predicts when a computer server will run out of memory.

Sound like pipe dream? IBM will try to prove it isn't by "eating its own dog food." Palmisano says it will take five to seven years to transform IBM's internal computers into systems that can be tapped when needed. The savings, he says, will be huge. "This initiative is worth a 5% productivity improvement each year for the next 5 to 10 years." That could translate into $2 billion to $3 billion a year for IBM.

Palmisano may be striking at the right moment. With IBM doing relatively well--profits from current operations were flat last quarter, at $1.7 billion--it's a good time to press its advantage. What's more, corporate buyers are looking for a way to save money and get away from the headache of managing complex computer systems. "The time is right to lay out an aggressive road map," says John A. McKinley, Merrill Lynch & Co.'s chief technology officer. That's why IBM plans to spend up to $800 million in the next year on a campaign to convince customers of the value of on-demand computing.

But for all the hoopla, don't expect that to happen overnight--or make much of a difference to IBM's top line in the near term. Many of the products, services, and industry standards that would make on-demand computing possible are still on the drawing board. "We're at least five years away before the technology starts going mainstream," says IDC analyst Melanie Posey. Even if IBM did have the goods, penny-pinching corporate buyers are wary of lofty promises.

Just ask Bob Egan a vice-president at Boise Cascade Corp. (BCC ), the $7 billion papermaker. As an IBM customer that processes $4.5 billion worth of e-commerce transactions every year, Boise should be an ideal customer. But Egan isn't buying into Palmisano's vision. "I can't imagine...they will be able to install a system that is cheaper" and more efficient, he says.

Palmisano may have IBM on the right track with on-demand computing. But a big payoff is not a sure thing.

By Spencer E. Ante in New York

    Before it's here, it's on the Bloomberg Terminal.