November 30, 2004
Looking Inside Intel: Missteps Force Strategic Change ‘Error of Judgment. Mea Culpa. I Learned a Lesson.’
One sign that Intel is having trouble dancing to technology's current beat may be the world's most expensive disco ball.
For a company holiday party next month, a handful of engineers had assembled a disco ball with hundreds of small reflective devices to hang above the dance floor. The mirrors are leftover projection- television chips from Intel's planned effort to enter the digital television market. Intel recently abandoned the project only 10 months after a splashy introduction at the Consumer Electronics Show in January.
For two decades, Intel has been the most sure-footed of Silicon Valley companies. But lately, it seems to have lost its way.
"They have made many wrong decisions and now its time for soul- searching and structural, not cosmetic, changes," said Ashok Kumar, a financial analyst at Raymond James & Associates.
This all portends an interesting inauguration for Intel's president, Paul Otellini, the longtime Intel marketing executive tapped by the board this month to become only the fourth chief executive in the company's history.
Otellini, who is 50, does not officially take the job until May. But next week, in one of his first official acts as the designated chief executive, he plans to present his strategy to Wall Street analysts. He may have a lot to answer for, including the 25 percent decline in Intel's stock price this year.
Otellini will tell analysts that he plans to focus on four areas for growth: international markets for desktop personal computers, mobile and wireless applications, the digital home, as well as a new initiative aimed at large corporate computing markets that Intel is calling the Digital Office.
The strategy is a significant shift a "right-hand turn," as Otellini likes to say from Intel's long-term obsession with making ever-faster computer chips. Instead, the company is now concentrating on what he calls platforms: complete systems aimed at both computing and consumer electronics markets.
Otellini insists that the recent missteps, including the premature introduction he himself made of the digital project, are simply a result of overly optimistic marketing.
"What was wrong was that I made the decision to go public on it at the Consumer Electronics Show," he said in a recent interview in Intel's headquarters in Santa Clara. "Error of judgment. Mea culpa. I learned a lesson."
The decision to preannounce an unproven technology was an uncharacteristic one for Intel, said G. Dan Hutcheson, president of VLSI Research and a longtime observer of the company. However, he said, it has been Otellini's ascendancy at the company that has changed the way it markets technology.
"As he came into power, Intel tried to become a more aggressive marketing company," he said. "They never seemingly made mistakes before and that was simply because they didn't preannounce. This is the classic failure of a company where the marketing guys are pushing the manufacturing guys more than what's there."
Intel is still a global leader in semiconductors, with an enviable international growth rate. But some of the company's marketing problems may become more acute before they are resolved. Until recently, selling Intel chips was easy: faster was better. Now, Otellini said, Intel intends to play the same game with the number of processor cores that can be embedded on a chip.
The hope is that by breaking problems into parts that can be computed by separate cores simultaneously, chips will continue to offer better performance. The problem with the strategy is that so far Intel is trailing AMD, IBM and Sun Microsystems, which all have their own aggressive multicore chip strategies.
Yet Intel's challenge in entering new markets also runs deeper, according to an engineer who worked on the ill-fated digital television project.
The engineer said that the company's failure to perfect the complex technology, known as liquid crystal on silicon, or LCOS, came from its inability to think beyond its expertise in manufacturing digital circuits.
He described sitting in meetings where the company's simulation models, based on its existing manufacturing process techniques, showed that 95 percent of the chips from each test wafer would be usable, while actual yields were closer to 4 percent.
LCOS may not be the last of Otellini's tough decisions.
"He has a tough inheritance," said Richard Doherty, a computer industry consultant who is president of Envisioneering.
"He has to take a cold hard look at these new markets, particularly the ones that aren't carrying their own weight."