"This is the third transition in Intel: from semiconductors to microprcessors, microprocessors to platform, to being an internet building-block supplier"

In September 2001, Intel appointed its first Chief Technology Officer in 33 years of existence, Patrick Gelsinger. Geeks everywhere would love Gelsinger: he gets his jollies from tech (he chuckled wickedly as he led this correspondent through a future that was 'IP everything'); he joined Intel in 1979 when he was just 18 years old; and he holds the record for being the youngest ever Group Vice President at the company (he made it at 32). On a recent visit to India, Gelsinger spoke to BT's R. Sukumar on his job, the unending search for the future at Intel, and the Intel India Development Centre at Bangalore. Excerpts.

You are Intel's first CTO. Why did Intel suddenly feel the need for a CTO?

Intel had CTOs in the past, just that they didn't call them that. There were people like Bob Noyce, Gordon Moore, Andy Grove... So I think we really had people who were the technical lighthouse for the company-setting a direction and really driving those. The other reason why having a CTO is appropriate now is that Intel has gone from being a fairly narrow company to being a very broad company.

I'm trying to build end-client devices for PCs and handhelds, trying to deliver the infrastructure for data centres, both computing and communications, I am trying to deliver the building blocks for the entire communication network. (Intel) has gone from being very narrow in computing to very broad in the whole range of internet communications and computing in the future.

As we've got very broad, having somebody whose job it is to try to put all those pieces together can become increasingly important. The various business groups are each focussed on their individual areas; my job is to tie the pieces together. Also (to) set a long-range common strategy and direction for our research. And try and work with industry and get Intel's strategies and direction to align with industry and international standards bodies.

You've been with Intel for over 20 years. How many times in this period have you seen the company reinvent itself?

The first reinvention of Intel was occurring just as I was getting into the company, where we really went from being a semiconductor company to being a microprocessor company. And calling it a microprocessor company in the early eighties was very appropriate. The real heart of the company was microprocessors.

Second real transformation was the transition into a platform company. PCI (Peripheral Components Interconnect, a standard used to link the personal computer to anything from modems and video cards) was the first major platform standard that we drove as a company. That moved us out of just worrying about our piece of the computer to worrying about the whole platform.

Now this is the third major transition in the company since I've been there: from semiconductors to microprocessors, microprocessors to platforms, now to being an internet building-block supplier.

If you have to define Intel's business in one line today how would you do it?

The definition we use in our mission statement is ''worldwide supplier of internet building blocks''. In some ways that may look limiting because communications is bigger than internet. The reason I don't think so, and we have a simple phrase we use in Intel: Everything over IP (Internet Protocol); IP everywhere. For instance, we have switched products (switches, etc) today; the future of all of that is going IP; future is Voice over IP; the future is entertainment over IP.

There are four networks that will collapse into one. You have the entertainment network, largely cable; the data network, internet; you have the telephony network; and you have the wireless network. We see all four moving to an IP network in the future, and that becomes the common denominator. That's why I don't think it (our mission statement) is all that limiting.

All tech companies are paranoid about being unable to keep pace with technology. What does it take to keep a company like Intel forward-focused?

At the simplest level, it is the mindset that 'technology wins'. Let me give you an example. When we moved from the 486 (chip) to the Pentium (chip), it was a bad business decision. Today, you can say, ''Were you nuts? Pentium went on to become a huge product line.'' But when we made that decision in 1993 to convert (from the 486 to the Pentium), we lost money in converting to the new technology. 486 chips were more profitable than Pentium chips. But we knew that Moore's Law (The number of transistors per square inch on integrated circuits will double approximately every 18 months, or the colloquial, processor power will double every 18 months) would win; that new technology would win.

''Not Another Bell Labs''

Last year we spent $3.8 billion on R&D. this year, our forecast is $4.1 billion. Now, in that you have research, development, and advanced development. We also follow the distributed lab model. Many of the labs for Intel don't work for me. I'm responsible for the overall direction of the company and those labs, but many of them don't work for me. We have about 7,000 non-product R&D personnel out of a total of close to 30,000 engineers. We want it to be very diffused. We're not trying to build Bell Labs. We're not trying to build Xerox PARC. Those were some of the finest research institutes but also the most ineffective... None of that (research) ended up affecting the company. I want to create a world class research and advanced development capability but more important to me is making it the most effective research and development capability. So many of the labs are either geographically dispersed, or organisationally dispersed, so they're next to the business groups we want to influence. What's a CTO? There are three different ends of the spectrum. One is Chief Research Officer. That's the IBM model. Paul Horn (Senior Vice President & Head Of Research) of IBM doesn't really have anything to do with the business. Then you can go to the other extreme-someone like Bill Raduchel (Executive Vice President & CTO) at AOL. He is heavily involved in the business decisions. The other aspect is, do you have resources? Someone like a Bill... he's a lone ranger. He has maybe 20-30 people. Whereas Paul has 5,000 people. The model we are trying to follow is somewhere in the middle. I own the research direction for Intel, I have labs that work for me, and I am very involved in the strategy of the company.

You have to say technology wins and let that drive your product, business, and strategic decisions. If you ever don't do that, obsolescence becomes your enemy. You must make obsolescence your friend by always committing to put your product in the front-end of the technology transition, even when you lose money, lose margins, in the process of doing so. If we don't capture technology transitions then we open ourselves up for our competitors to do so.

We have to, then, cast a very broad net, because you don't innovate all the technologies. I've got nearly a thousand people doing innovative work inside Intel, but there are hundreds of thousands of people outside of Intel doing innovative work. We've structured our research programme into what I call research without walls. The idea behind that is I partner very broadly with universities, with other companies, so that we have a finger on the pulse of all these other innovations. Let's take 802.11 (a wireless networking technology) as an example; Intel wasn't the first with it; Atheros was ahead of us. We weren't terribly behind, but we weren't the first. It is a big thrust for Intel now, and we are positioned to catch up very quickly. But if we didn't have a finger on the pulse of the industry, if we were late, we'd never be able to participate in a very critical technology transition.

So, we do that through our research. We do that through the research of other companies. And we do that through our venture plan.

There are two major areas where you can spot innovation. One is in the academic community; the other is in the venture community. So, between the research stuff that I own, the academic relations part, and the venture stuff that Leslie L. Valdasz, Sr. Vice President, Intel Capital, owns-we partner very heavily on the very early stage deals-(we cover everything).

Intel runs a development centre, the Intel India Development Centre, in Bangalore. How does the IIDC fit in the company's larger research plans?

Let's think about it on two dimensions. One, we aim at creating a larger international R&D capacity. The three focus countries for that are Russia, India, and China. So, if you want to know where Intel is going to increase its research efforts, it is in those areas. We like other countries as well, but are proactively putting in place plans to increase our presence in those (three). There are two reasons for that. They're emerging pools of talent, typically lower cost talent; and also emerging markets as well-the three are the fast-growing large markets of the future. So, the more we are in there, the more influential we become in those markets-the more local we are.

Another aspect to look at is that it takes a while to mature a new development centre. Initially, a new development centre becomes largely a job shop: do this, and when you are done, give me the results, and then I'll give you another job. And then they become more mature, and start taking on not jobs, but projects. And they become even more mature, and then it's not projects, but products. Products become strategies.

IIDC is just a couple of years old. We've gone from being a job shop and are just getting to the point now where we get into projects. Our plan is to mature it to the point where it is taking on entire product lines for the company. That it becomes a very important part of our overall development capacity over time.

It's doing well. We're pretty happy. It is now the largest area for e-business development for the company. We've more people working in that area there than at anyplace else at this point. It is taking on an increasing role in our communications product-line.

Where does an Intel CTO go from here?

I don't know. It's not as if we've had CTOs before. But CTOs tend to stay CTOs for a long time.