Intel CEO Predicts 'A Very Strong Second Half'
Page 1 of 1
|Intel CEO Paul Otellini speaks at the Web 2.0 Summit|
Photo: David Needle
"We don't see the end of silicon for quite some time, but it will happen," Intel's CEO Paul Otellini said in an onstage interview here at the Web 2.0 Summit that touched on a variety of topics.
One of those included some tantalizing hints on where Intel sees its core technology heading.
"We have prototypes that are not silicon-based," Otellini said, while declining to give any more detail other than, "it's cool." But he pointed out that silicon has plenty of life left in it -- and that Intel should know, since it's built transistor structures that are three generations from coming to market, with each product generation potentially lasting two to four years.
As for the more near-term PC and portable computer market, Otellini is bullish, particularly on the consumer side for this year.
But starting next year, he expects enterprises to ramp up buying of new servers as well as PCs for a number of reasons.
"Corporate CFOs have been putting the clamps on spending, but I expect that to change in 2010" as the economy improves, Otellini said.
If, as he expects, the economy continues to rebound, CFOs will have the confidence to authorize more purchases. In Otellini's view, it's a no-brainer because the average corporate desktop and notebook are five and four years old, respectively, and badly in need of replacement.
"They have to be refreshed because they don't run Windows 7, dont have the same level of securability and manageability, and it costs more to maintain them than replace them," he said. "CFOs buy that argument."
"The machines we're deploying this year are so much better," he added. "They require less energy, have a smaller footprint, greater capacity and their performance is better, so that's driving server sales."
Continually producing chips that are smaller, more powerful and more energy-efficient requires hefty investment for Intel. Asked about its chip manufacturing facilities, Otellini said Intel has invested $7 billion in plants focused on the latest 32-nanometer technology "that should last us the next couple of years."
He said Intel plans to make a decision on where and when to invest in 22nm six months from now.
When looking where to build a new plant or fab from scratch, Otellini added the U.S. is the least attractive option.
"It's about a billion-dollar difference because of government incentives and tax rates, because [other countries] want those facilities in their backyards. The U.S. doesn't have that kind of industrial policy," he said.
Cloud computing, netbooks, and thinking healthy at Intel
Otellini touched briefly on several other topical issues. He noted that netbook sales are on a faster pace than the popular Nintendo Wii videogame system or the Apple iPhone.
"We're going to put more capability in netbooks to make them more like smartphones with GPS features," he said. He added that netbooks, by virtue of their larger screen, are more practical for many applications than a smartphone.
"If you want your kid to do homework, doing research on the iPhone is not the best device," he said.
Another key trend -- cloud computing -- is widely seen as driving a major shift in the software business. But Otellini said that while it's a clear force in the IT market, it doesn't change Intel's plans significantly from a hardware perspective "except we'll sell a lot more server chips."
"The difference is some of the usage models," he added. "Pushing some applications out to the cloud makes sense [but] I don't think I'm going to be putting my payroll or accounts receivable in someone else's cloud."
Shifting gears, he also touted wellness programs that Intel instituted a few years ago.
"We've taken our costs down by making sure people are more fit and tested," he said. "We need more of this as a society because prevention is a better long term investment."
Computing like Star Trek
During a brief Q&A session, Otellini was asked when mainstream computing would no longer require keyboards.
"Cognitive speech recognition has been about five years away for the past thirty years and I think it's still another five years away," he said, drawing a good round of laughter from the audience. "It takes a lot of intelligence to make it really simple. You have to be Star Trek perfect and we are so far away."
That said, he noted that advances like true, consistent speech recognition requires the kind of processing breakthroughs Intel thrives on -- and that bodes well for sale of advanced chips down the road.
"There's an insatiable demand," he said.