At Intel, Speed Isn't Everything
(FORTUNE Magazine) – The world's leading chipmaker has been obsessed with building faster and faster microprocessors. And it still is. But with CEO Craig Barrett required by company bylaws to retire next year, his designated successor president, Paul Otellini, a former marketer, has been pushing Intel to listen to what else clients want and develop products accordingly. Shortly after Intel reported record sales and better-than-expected profits for the fourth quarter, we caught up with Otellini to talk about tech spending, Moore's Law, and Atkins. --David Kirkpatrick
So businesses are spending on tech again?
They never stopped spending. They shifted from hardware to software and services. Now we're seeing a shift back toward hardware.
You're presumably going to be CEO soon. Is it scary to think about being the first non-engineer to head Intel?
I consider myself to be a product guy even though I'm not an engineer. I've had my fingerprints on our products for almost 30 years. Plus it allows you to ask the higher-level dumb questions. Why can't we do X? Why isn't it possible? We do not put a feature into a microprocessor anymore unless we know how we will market it and what the end-user benefits will be. That is 180 degrees from the "build a better mousetrap and world will beat a path to our door" model that we had for a long time. We've changed in the past three to four years.
We look at markets more holistically now. Take our new Centrino mobile technology--it wasn't sufficient to build the world's best microprocessor for a laptop. We went further and designed a group of chips for the fastest, lightest computers with the longest battery life. We have a similarly encompassing vision for the digital home. We are taking an end-to-end view and designing our future products to deliver a much more seamless, integrated experience.
Moore's Law, formulated almost 40 years ago by Intel chairman emeritus Gordon Moore, says that the number of transistors on a silicon chip doubles every couple of years. How long will it continue?
We have already built prototypes for five generations of transistor structures beyond our 90 nanometer ones now. That will take us out to 2011 or so. The 2011 transistor structure is smaller than the diameter of human DNA. Beyond that at some point you start hitting some fundamental limits, but Gordon always said he could never see how things would be implemented more than five years out anyway.
Does Intel have a unique responsibility now that you produce almost all the world's computers?
If anything there's a responsibility to advance the puck. Technology stagnates if someone doesn't invest. If we stop where we are, the state of information technology and computing will stagnate.
Now that 70% of revenues come from outside the U.S., is Intel still an American company?
We are headquartered here, and our allegiance is here, but our people and our new opportunities are increasingly global. Our growth has been largely in emerging markets over the past couple years. We are now populating tier-two and -three cities in China. These are cities you cannot find on most maps.
But will most of your production remain in the U.S.?
I would think so, because the most efficient way for us to build a new fab is to put it adjacent to an existing one. The next three coming on line are in New Mexico, Oregon, and Ireland. On the other hand these factories are getting so expensive that incentives can be a big part of the decision. And essentially there are no incentives in the U.S. anymore.
Last question: How much weight have you lost?
Twenty-eight pounds. It was Atkins. The hardest part, as an Italian, is giving up pasta.