Craig R Barrett, President and CEO, Intel Corp.
Dr Craig Barrett took over from
the legendary Andy Grove as the President and CEO in May 1998, when Intel was facing
turbulent times in its core microprocessor business. Moving away from Grove’s shadow, and
creating his own identity, Barrett is clearly broadening the focus of Intel to take into
its fold aspects like networking, processors for high-end systems and driving the
development and deployment of applications on the Intel platform. In short, Barrett is
trying to don the cap of an IT evangelist and establish Intel as the leader of the IT
industry. Not that the industry doesn’t need either of them.
As part of his evangelizing
efforts, the 59-year-old Barrett has been coming to India regularly for the past three
years. This time, he came as President and CEO of the company. He met the Prime Minister’s
IT Task Force and addressed the Indian IT industry and users. In an extensive interview to
DATAQUEST, Barrett spoke about his vision and Intel’s stature in the current spate of
things. Read on…
You have taken over as the CEO of
Intel after a long innings from Andy Grove. What is your vision and in what way is your
outlook different from Grove as far as Intel is concerned?
I don’t see any significant differences with Andy. And first of all, Andy is not out. He
is still our very active Chairman. And I have been working very closely with him for the
last ten years. And I think our thoughts are very similar to each other. The biggest
challenge I currently face is one of growth. In the last ten years we grew very rapidly,
with average growth rates of over 30%. It’s very difficult to grow at large growth rates
when you are a big company. And fortunately in the last quarter, we have been able to grow
very well after a year and a half, which was 1997 and most of 1998. Getting back on the
growth track, both from the revenue and the earnings standpoint, is the biggest challenge
that the company faces today.
Talking about growth, the major
challenge that Intel faces is in the lower-end segment, the sub-$1,000 PC segment, where
the marketshare of the company has come down from around 90% to 45%. What are the
company’s plans to regain this lost marketshare?
Well, you need to look at Intel’s plans from a larger perspective. First of all, we intend
to be very aggressive in the sub-$1,000 PC category. If you look at anyone’s assessment of
the latest additions to the Celeron processor, the 128k, embedded cache etcetera then you
would find very competitive price/performance. And it will do very well in that space. The
Pentium II family is doing very well, and we have a substantial opportunity for growth in
the higher-end workstations and server marketplace, from both the high performance 32-bit
and 64-bit architectures.
So, if you look at the opportunity for
growth, then I want to be competitive in the mobile space, low-end space, performance
desktop, workstation and server spaces. You have to integrate all that and look at our
growth opportunity. Our competition has chosen to concentrate on one of those segments.
And you can see how effective they are if you look at their profit and loss statement. AMD
has been there for the last ten years. National/Cyrix has been there for the past 15
years. National recently witnessed what was a ten-year low in their stock price.
Are you saying that it may not be
able to make money at its current pricing levels, etc. And is that the reason that Intel
is coming under pressure?
I am only stating the obvious fact that they are not making money.
Recently, after your financial
results were announced, Executive VP Paul Otellini said, “Our segmentation strategy
is working. The high-end growth is offsetting erosion in the lower-end.” Does this
imply a tacit acceptance of the fact that Intel’s near-monopoly position is under severe
threat and the company may have to vacate this space in the future?
Component manufacturers, OEMs and consumers
agree on one thing-selling PCs in the $500-$1000 price range is a good deal only for the
consumers. It is not a good deal for the OEM PC manufacturers or the component
manufacturers. Therefore, if you are going to have that sort of segmentation occur in the
marketplace, then what you would like to do is isolate it to the extent you can from the
rest of your business. That’s what the segmentation strategy is. We have a family to
compete at each segment. The Celeron family of processors competes in the low-end space.
The Pentium II family is in the next level. Then you have Xeon, Merced etcetera. The
recognition is that some places you get paid for a performance premium, and some places
you will not get paid at all, and segmenting your products and brands into different
categories aids in recuperating your investments.
Does this imply that Intel’s future
investments will be directed at the upper end efforts, in terms of R&D, marketing
efforts and everything else that goes into creating a product?
Well, we intend to be competitive in each one of those segments. But clearly, you would
tailor your investments to where you get your greatest return from. We are clearly making
a big investment in the 64-bit family of processors. We don’t have a product in the market
today. So that is totally investment and no return. The Pentium II-Xeon family is doing
very well in the high end. So, we will continue to invest in all the categories and be
very technically competitive in all those categories. But I think there is more
opportunity to grow in the higher end than in the lower end.
Closer home, specifically in Asia,
the markets are price-sensitive to a larger extent. Would you be following a different
marketing strategy because the pressure on you might be higher, say, than in the US?
You know I keep hearing that the Asia market is more price-sensitive than the rest; the
Latin America market is more price-sensitive than the rest; the European market is more
price-sensitive than the rest; and the US market, where the sub-$1,000 PC had its origin,
is more price-sensitive than the rest. So I don’t agree that the premise of your question
is correct. The premise, you know, is the same everywhere.
And I don’t think it is necessarily bad
that they [competition] have the [sub-$1,000] segment of the business. With the sub-$1,000
category, there are a number of things. One is it gets more people interested in computing
because more people will buy computers. That’s where the industry will focus on. Second,
it has shot Larry Ellison [CEO, Oracle Corp] in his foot. His network computer, which was
going to change the face of computing, and suddenly you have the sub-$1,000 PC which
competes with the NC. And nobody asks me about the NC anymore. The marketplace has spoken.
NCs are replacing only dumb terminals, and not the PCs.
A recent survey by Ziff-Davis
pointed out that in a PC purchase decision, ‘Intel Inside’ was the #1 factor till a year
back, but now it has slipped down to #3. Does this imply that a greater emphasis is now on
the brand name of the machine than on the chip inside?
Not particularly. I think you have to look at where and how the survey was taken. If you
look at the corporate area, the ‘Intel Inside’ brand is very significant. The
competition’s focus is on almost exclusively the low-end retail marketplace. That category
(sub-$1,000) quite often is entry level, new participants, and price is a very important
feature there. These factors are more important than the OEM brand, the ‘Intel Inside’
brand and whatever. The US retail marketplace is very interesting to play, but it
constitutes less than 10% of the computer marketplace worldwide. And we are interested in
the totality of the computer marketplace.
We spend extraordinarily large sums of
money in terms of our direct advertising and corporate advertising (Intel Inside). We
would be terribly foolish to spend the money if the OEM name was the only thing that drove
a consumer into purchasing the PC…
…And you don’t see this trend
changing in the future?
There is very strong end-user recognition and commitment to ‘Intel Inside’. And
we don’t see that changing in the future from the statistical studies that we do. If it is
changing, then I will have to do one of the two things-either fire my marketing manager or
I do something else [i.e. not make PC processors].
Processors are encompassing more
and more capabilities like graphics, audio, video etc. What are Intel’s specific plans in
We are introducing the Katmai family of processors in Q1 1999 and it has 72 instructions,
basically capabilities like imaging, animation and such other calculations. So that is one
area. Input/output area too is an area which will get increased focus. Then there is
Merced, and the future processors, like McKinley, Willamette and so on and so forth.
Processor speeds are not going out of vogue. More and more speed will continue to be
needed for the next 15-20 years. We will double the processing speeds every 18 months. So
this will continue, and future processors will have added instructions, features for new
data types like speech recognition or increased graphics, animation capability, 64-bit for
long addresses and other things we are going to address. They are quite boring, mundane,
but that’s what keeps us going.
But what about some of the
fundamental changes that you will find in a chip. One is obviously speed, but what about
other integration on the chip?
Well, you can put more and more transistors in your processors. More transistors
mean more capabilities. You can put the graphic function, memory control function etc. If
you want to do a lower-cost solution you typically put in more functionality into the
chip. You see some movement in the server space, where performance is driven by access to
the local cache memory. Embedded cache is incorporated into the processor, especially in
the server workstation space, where it is critical. So you see solutions which come out
like one Meg [MB] and two Meg cache memory or more in the future incorporated into the
processor itself. The other area which could be important in the future is general area of
security. With the internet, security and privacy are becoming of paramount importance.
Security is more of software architecture, but you could also incorporate security aspects
into hardware architecture. So in each one of those areas you can put in so much
additional functionality into the chip.
A delay in Merced means
consolidation on the Alpha front, especially with Compaq behind it. How do you read the
situation in this light?
You know I get a real kick out of talking about Alpha. Compaq is Intel’s biggest IA-64
server customer. And that’s where the profits will come for Compaq and not sub-$1,000 PCs.
They buy Digital and what comes with it is the Alpha business worth $4 billion. And the
media is excited about the fact that Eckhard Pfeiffer comes out and says, “We are
going to back Alpha.” What could he say? No, we are not gonna back it? I don’t care
about the $4-billion business that we have today based on Alpha. He has to say that he’s
gonna back it.
There is only a six-month delay in Merced. And there is a whole range of processor
families that we have announced in the Microprocessor Forum. There is Merced, McKinley,
Cascades, Foster and whatever, whatever they are. We are doing parallel designs in the
32-bit family as well as the 64-bit family. Every computer OEM, including Digital before
Compaq acquisition and Compaq after it acquired Digital, has committed to the Merced
family and the IA-64 architecture. There was a postmortem done in the Forum about the
existing RISC architectures in competition with Intel architecture. Basically, there is
one computer company that hasn’t committed to the Intel architecture. You know who it
Yeah, Sun. They have said that they will use the Intel architecture and port
their software on it. Put their software on Intel architecture rather than incorporate
Intel architecture into their boxes. We are bullish and confident on the 64-bit family and
the high-performance 32-bit family. The two will coexist for a long period of time. The
32-bit server family gives you high performance today, and competes effectively with
PowerPC, Sparc, Alpha and MIPS. The Pentium II-Xeon family compares very well with them
both from an absolute performance standpoint and the price/performance standpoint.
Going India-specific, if you had to
set up a manufacturing facility, what would be your imperatives which would influence the
The answer is very simple, and this answer is what I give to every country. I suggest them
to become internationally competitive, and I would be putting a plant anywhere in the
world. Our business is one of international competition. We may have a manufacturing plant
in India but it would have to produce far more than India could ever consume. So people
will have to be competitive from an exports view. So all the elements that go into
determining whether a plant is competitive with any other international location for
exports is exactly what has to be satisfied by India. It includes telecommunications
infrastructure, transport infrastructure, workforce, taxes, government attitude, utility
infrastructure and surrounding industry infrastructure-if Intel puts up a wafer
fabrication facility, there has to be applied materials, chemical industry etc. I think it
is a long way off.
But I think the opportunity is much greater
if Intel acts as a venture capitalist here. As we do in the Western world, we can invest
in small Indian companies with exciting new technology in and around the computer
industry. I think we would qualify as the biggest venture capitalist in the world in this
area. Last year, we made over 100 investments worth $350 million in small equity
investments in small start-up companies. I want to extend that program in India. To me,
that’s a much stronger approach to getting involved in the Indian economy and value add to
the country. I could probably put one manufacturing plant in India, but this way I could
make tens of investments in small companies, which would be the foundation of your future
IT business. For me that’s much more exciting possibility. And we are putting in
infrastructure to support that initiative.
Considering the manpower-related
skills which India has, do you think hardware design is a feasible area of investment?
Sure. That’s entirely possible. We have design teams working in Israel and Malaysia, but
both might be working on the same product. But, right now our major focus is working with
your software industry, rather than hardware design. You know, value-added software
content, digital content creation. Those areas I think are exciting for the future.
You have been visiting India
regularly for the past three years now. What is the objective behind coming to India every
year? What are you looking at?
We have formally been here for ten years. Our objective is to bring the latest technology
to Indian PC manufacturers. I come and talk to our customers. I come and look at the
marketplace and how computers are sold, what’s exciting in the marketplace, the
distribution channels-which of them work and which of them don’t. I talk to government
leaders, as I did this time [the Task Force], about what takes place in India, what you
have to do to become a major player in the IT industry and what you have to do with the
infrastructure whether it is telecommunications, the internet, software, business startups
and taxation, and get an overall feel about the industry here. I am little bit of a
hands-on person. I like to come and see what it looks like rather than read about it in
the country manager’s progress reports. And I just like to visit India.
You have recently invested in a
Linux company, Red Hat Software Inc. How do you perceive the potential of Linux vis-a-vis
Windows NT? Do you see Lintel instead of Wintel in the coming years?
Our strategy is very simple and it has been that we would like Intel architecture to be
the port for any OS. Linux is getting increasing popularity in a couple of categories-ISPs
and an increasing number of corporations are embracing Linux. So we support Linux, just
like we support other Unix variants. We support Windows. We support anybody who wants to
port their OS onto our hardware. Sometimes we make investments and some other times we
don’t have to make investments. Linux is a free software. This investment was done to
support them to be viable. And frankly Unix is the one when you get to those big
enterprises, in terms of scalability, reliability. You know Unix still has a lot of
popularity when you talk about enterprise applications. The media talks about Wintel all
the time. We don’t use the phrase. It is Intel with any OS.
and SIVAKUMAR V,
in New Delhi.