Analysis: Intel's big push for convergence

Chip giant Intel has long been hoping to make its computer processing expertise pay off in the communications and networking arenas but reports that it is seeking to sell several comms-related businesses suggest the strategy has stumbled. silicon.com sister site CNET News.com's Tom Krazit investigates.

Intel has been pushing "convergence" since the height of the dot-com era. Seven years later, it's looking a little more like the new word of the day is "divergence".

Intel might be contemplating the sale of several communications-related businesses, according to recent reports from the San Jose Mercury News and The Wall Street Journal. In April, Intel CEO Paul Otellini said the company was undertaking a strategic review of its operations to cut costs. According to the latest reports, that could involve a restructuring which would see Intel dumping several products and businesses that have failed to live up to expectations and are unrelated to PC and server processors.

An Intel representative declined to respond to the reports, saying the company will not comment on any possible actions until it has completed the review Otellini discussed.

At one point, Intel had envisioned itself getting into several communications businesses, such as networking and cellular, as a way to continue its strong growth of the 1990s despite the slowing growth of the PC market. In a vision it called "convergence", the company wanted to combine its expertise in computer processing with newly developed talents in communications to create chips for PCs, phones and other devices that could talk to each other.

Some of that vision has become reality but several parts have not panned out. Intel's bets on wi-fi chips coupled to its Pentium M processors paid off handsomely with the success of the Centrino laptop package of chips. But similar efforts to develop a combination applications-processor and cellular-communication chip for mobile phones have barely made a dent in Texas Instruments' control of that business. And Intel's networking-processor business has not turned into a major money-spinner either.

In December 2003, Intel admitted its strategy was not working out as it had hoped. It announced plans to combine its separate wireless units into one business and take a $600m charge. Since then, Intel's main business has been besieged on multiple fronts and seen renewed pressure from AMD. But extracting itself from the communications arena isn't a simple matter of slicing off a division, say analysts.

Intel may not make a ton of money from every business it operates but it can learn a great deal by experimenting with silicon in places outside of its core markets, said Jim McGregor, an analyst with In-Stat MDR. For example, the company hasn't got rich selling network processors and boards but it has learned how to build chips that can run reliably for long periods in demanding environments, and some of that knowledge can help designers of its server products, he said.

But it has been hard for Intel to transfer its expertise from computing to communications, according to Forward Concepts analyst Will Strauss. The two disciplines require different ways of approaching circuit design but at one point Intel thought it could just use its manufacturing might to out-build the competition, he said.

If Intel has decided to put the some of the communications businesses on the table, its new challenge is figuring out the next big thing that will drive the computing world forward, In-Stat's McGregor said. In recent years, the company has focused on increasing its share of the components within a PC or server under Otellini's platform strategy.

It has also made aggressive moves into the digital home, promoting PCs and components as the heart of the future living room. But the underlying situation that led Intel into the communications arena hasn't changed: PC sales are still forecasted to slow down over the next five years.

McGregor said: "If it's not communications, what is it? Do they focus on their core competencies, or are they going to be a technology driver 10 years from now?"

Tom Krazit writes for CNET News.com