Intel CEO Craig Barrett: says it's time for the United States to wake up to the new global realities about technology and economics.
To keep Intel at the forefront of the microprocessor business, Intel's CEO poured cash into research and development while institutional investors called for layoffs. That perseverance seems to have paid off. With the tech economy gradually gaining steam, Intel beat expectations and turned a healthy profit during the third quarter.
Craig Barrett, CEO, Intel
Now, Barrett, who spoke Tuesday at an industry conference sponsored by Gartner, is out to prove that the convergence of computing, communications and digital content will help drive an industry resurgence.
Barrett's speech was also a call to arms. Intel's CEO sees a faulty educational system that does little to teach adequate math and science skills as a major problem the high-tech economy--and the nation--will need to address in the coming years. Among other ideas, Barrett advocates giving foreign nationals who are recent engineering school graduates instant green cards in order to shore up the domestic engineering talent pool.
While most of the world sees high tech as the leading driver of the economy in the coming years, the United States--and in particular Intel's home state of California--treats the industry as a second-class citizen, according to Barrett.
Coupled with the growth of Intel's business outside of the United States, that's led the company to place a greater value on its overseas operations.
Barrett sat down with CNET News.com editors to discuss global economic issues, along with how Intel will deliver new technologies and why companies should adopt them.Q: What will Intel have in store to coincide with Microsoft's introduction of Longhorn, its next version of Windows?
How much of a sales driver for Intel will Longhorn be?
Kind of tough to say. You are at least a couple of years away from delivery and a lot can happen.
To put it another way, following Moore's Law, what kind of horsepower would be available if you bought a new machine at the same time that the new OS is out?
I think you will see a variety of technologies provided in addition to the standard more clocks or cycles. I think you will see the integration of technologies, whether it's wireless or some degree of virtualization, such as Vanderpool. You will most likely see movement probably beyond hyperthreading into the multiple core situations, which are starting at a high level and are cascading down to the desktop.
Will there be enough horsepower at that time for speech recognition?
I'm not sure there is ever enough for speech recognition. We've been talking about speech recognition for 20 years, since Intel built the first digital signal processor. Speech recognition was always the next application, for next year. It's clearly getting better and better, and you can do all sorts of memos--and Andy Grove uses it. So it's getting there. But you assume Moore's Law, processing power, continues to grow, so we are going to double every 18 months or so. In four years you get four times as much as today to do that sort of digital signal processing.
When you are working with Microsoft, do they have a reference platform for the minimum requirements for a 2005 or 2006 system?
Obviously, we share road maps with Microsoft and other people. That benchmark is out there for them. We both have to work in the multiyear lead-time situations. It may take us three years to go from design to bottoms-up product out the door.
What kind of an impact will Vanderpool, which allows users to partition individual chips inside their computers, have when it comes to the marketplace?
|I don't want to pick on Bush. I don't think the Clinton administration had a good handle on this either.|
Will the consumer market or the business market be the dominant market for Intel in the coming years?
I think that the theme is convergence, and it is equally dominant on both sides.
Is there a point at which the security tax is too high?
We're a pretty homogeneous environment at Intel. We're probably as big of a target as any in the world. It's not an unbearable burden for us to carry at this stage. Would I wish to get more net benefit out of this capability to overcome that tax? Yes. Would I stop deployment of new technology? No. Do I like it? No, but we are still learning.
How long before we see the Trusted Computing platform supported on multiple systems?
I think it will take longer than most people anticipate. Maybe in the five- to 10-year range.
What do you say to people who say Intel is out there pushing Moore's Law but have a 600MHz machine and say they are doing just fine?
If you look at the convergence of computing, communications and digital content, and you think 600MHz is enough, that's fine. I'll take my 3.2GHz machine over your 600MHz any day. If I'm into rich content, data mining and multitasking, (I need more).
What's your take on the telecoms? Obviously it's a core part of your business.
I think they are still struggling with two things: the movement to more standard building blocks that will lower the cost of telecommunications. The other thing is the competitive environment. That comes in two categories?-rules and regulations and the battle between the wireline and wireless guys and the cable guys. And you can add voice over IP. They are trying to deal with how to grow if the revenue per user goes down. You grow by adding new services, but you have to do that at the same time that people are coming in and trying to undermine your basic business model. At the same time, the government says that if you make an investment, you have to share. So they are kind of between a rock and a hard place.
Is Sun Microsystems still viable?
Scott (McNealy) has a parallel on the desktop, and that is Apple. They are sophisticated software companies who are wed to hardware revenues and margins. And they are both wed to proprietary hardware. The question is, what is the business model that you can carry over that is successful? Apple's model has been they're increasingly happy with 2 percent of the marketplace, and they are happy to cede the other 98 percent of the marketplace to other people.
I don't know if Scott is on that same track or not. He has had 10 quarters of decreasing revenues. He is increasingly facing competition from Itanium and the more standard building blocks and at the high end from big-iron companies. So he has to decide whether he is going to be a closed-source, proprietary-limited player, as Apple has decided to do, or he is going to do something different. He is at least using Intel architecture at the bottom of the scale. Customers want Solaris on cost-effective building blocks. We have tried to win designs at Sun, as we do at Apple all of the time. I don't have any unique advice for these guys.
Is Sun still a valuable partner, and will that relationship grow in the future?
I'm biased in this area. I think that SPARC is not a long-term player in this area. To be a long-term player, (McNealy) will have to start to accept product margins that make (Sun's) price/performance more attractive. We compete enough with them in the big-iron area to know what the price/performance advantage of our architecture is, and that's why he is losing. He will either take a huge margin hit, or he needs to do something different. But he has a lot of cash, and if his board supports him, he can continue to move at the speed he has been moving.
Will Intel ever be able to crack Apple?
|The government needs policies that promote investment, not inhibit investment.|
The OS X kernel runs just fine on Intel. Just a matter of the app stack to stick on top of that. But you'll have to talk to Steve about that. We just try to get design wins with these guys.
Intel does have some marketing approaches in common with Apple, such as the consumer approach with Centrino.
With Centrino, we're trying to make sure the consumer gets the whole brand experience. So we have to make sure it works with what's out there. People are going away from thinking about what particular protocol you have out there to a smart system that will pick out the best connection out there period, and the brand will migrate to that.
Is that something that you are going to provide as a feature on Intel-based systems?
There is a lot of work going on in tuneable radios and giving the user the best possible experience. We started off with 802.11b, but then it's a, g, x or whatever coming. So you want to come into the room and punch a button and say, "Give me the best connection, and the radio is going to tune to that." And hopefully, we will have the billing in place to handle it. That's clearly where the industry is going.
I thought you might ask me, "Hey, Craig, you are a U.S. citizen. Don't you have a patriotic duty to support the U.S. economy?" Increasingly our business, like other businesses, is international in nature. We will maintain a position as a U.S. flagship company. But we will do business around the globe and use resources around the world. Intel and Microsoft and Hewlett-Packard and Dell are increasingly in this mold. The world is the marketplace. We have a patriotic allegiance to the United States and we will speak as loudly as we can. But if people won't listen, that won't keep us from doing the right thing for our company.