The Intel insider
What Paul Otollini says and thinks will determine the speed--and price--of your next Intel processor.
Throughout that long career, the 51-year-old Otellini has played a part in many of the important events shaping the company. His various roles have ranged from managing relations with IBM in the early 1980s to running sales and marketing for the entire company. He oversaw Intel?s push into Ethernet silicon and PC chipsets and also served as technical assistant to former CEO Andy Grove.
Since 1998, Otellini has been in charge of the Intel Architecture Group, which means he is responsible for PC, workstation and server processors as well as research and development. And ever since the transfer in power last year that officially elevated Craig Barrett to CEO, Otellini has been frequently mentioned as one of the top candidates to be next in line for the No. 1 job at Intel.
But that's--possibly--the future and Otellini has more than enough on
this plate in the here and now. The challenges are increasing, what with
the drop-off in the economy in the last year and a half as well as
Intel's new mobile processors smaller, faster Paul Otellini, VP, Intel Architecture Group |
Otellini, who spends about 40 percent of his time on the road, recently sat down with CNET News.com to talk about the business and what's on tap in the near future as Intel seeks to spur more people to buy systems based on its P4 chip.
Q: You're out there in the trenches, so to speak. How does it look in
the PC market right now? How does it look for the second half?
Our view is that the industry should return to normal growth patterns in
the second half. For long-term growth in 2002 to 2007 or 2008, our view
is that it returns to the trend line. Historically the first half is
bigger than the second half.
Would you care to comment on the rivalry with AMD?
No.
What is it about processors that get people so riled up?
The rabidness tends to be a very small segment of the population. The
computer industry is 150 million units per year--and the number of
tech zealots is under 10 percent. You and I may happen to live in that
world, but it's like living in Washington, D.C. You think everyone is into
politics.
Can you compare 2001 to 1998, when you first stepped into your
current job and the semiconductor market also took a downturn?
I think it's almost an incomparable year. Depending on how the year
sorts out, it's very close to being the first no growth year in the
history of the PC. Based upon that, there's really been nothing in the
20 years it's been out that compares to this. I think about that a lot.
Is this the beginning of a trend or a reflection of the market we saw in
1999 and 2000? I think it was the latter. Demand was above the trend
line. It accelerated in 1999 and 2000. Some of that had to do with the
Y2K prepurchases and there was the Internet boom. If you take 1999,
2000 and 2001 together and divide by three you get a much more accurate
picture.
What made you decide to cut prices on Pentium 4 in April? Was it a
specific incident? Or was there a consensus inside Intel that price cuts
were needed to get the chip moving?
We cut the prices on every product every quarter. The April move was
very consistent with that. The more interesting view of that is what we
talked about in the (second-quarter earnings) analyst call where we
talked about being even more aggressive with Pentium 4 throughout the
rest of this year to essentially drive it into the mainstream.
So were those price cuts way overblown?
I do believe that the moves we made were a significant acceleration
versus our prior trends. We are going to push the Pentium 4 to
transition from the Pentium IIII, and we're driving the P4 down to
mainstream price points from $800 on up. That was driven by three
things: the health of the product, overall capacity and the introduction
of the 845 chipset ahead of schedule.
Does the state of the PC market worry you at all? Analysts say that
the majority of PCs are being purchased at prices that range from $800
to $1200 and that there is little activity at prices above $1500. Does
this mean that Intel will have to permanently change its processor
pricing structure to offer lower priced chips that fit into this band?
Or is this a function of the current market?
That's what Celeron was all about. When we introduced it, it was to
satisfy prices under $1000 and now it?s under $800. I think there's a
long-term kind of secular trend here, in terms of the electronics to
become increasingly integrated and (manufacturers) to pass on the cost
savings. Graphics, I think, is a very good example. Most of our chipsets
include integrated graphics now. For the low-end PCs, the very function
of integration takes $10 to $20 out of the bill of materials. That's
happening in every element of the computer. There's no need to rethink
pricing. Our entire business model is set up to displace our old
products with our new products.
Our entire business model is predicated on displacing our older product line with newer products. We have been doing this as a company for over 30 years and in the PC arena for 20 years. We will use our design, marketing and manufacturing assets to make the transition from Pentium III to Pentium 4 happen as quickly as possible.
With price cuts in effect and the 845 chipset shipping relatively
soon, do you feel Pentium 4 is about to come into its own?
It's already in major volumes. We said the volumes doubled again this
quarter. We'll just continue to grow over the year. The industry now is
very much a price point displacement model. If you hit the right price
point?people will always take better technology. The new platform
(Pentium 4) and a new OS (Windows XP) is incentive for people to
upgrade.
There's something like 50 million Pentium IIs out there under 500MHz and
another 50 million Pentium IIIs at 500MHz. You've got a potential for
upgrades that hasn't existed for a number of years.
Pentium 4 will hit 2GHz very soon. What do you expect people to do
with processors of that speed? What do you think when you hear people
say that you don't need that kind or processing power?
I have literally had that same question asked of me for the last 25
years. I think that's bunk. Would you be happy with your 286 or your 386
or your Pentium 100? No. The beauty of the industry is that the software
grows to consume the MIPS that are available. They demand more MIPS to
get similar performance as the previous generation. Ask anyone who's
doing advanced work--photography or ripping CDs or home videos--and
they'll tell you, 'Give me more MIPS.' I have a 20-year-old son. This is
how he spends his time. He's always bugging me for a faster computer.
What will the PC look like in five years, aside from faster
processors?
It will be smaller, faster and wirelessly connected. Always connected--there will be a seamless connection for the devices you use with it (such
as PDAs.)
What about servers? You must have breathed a sigh of relief when
Itanium finally "shipped" How important is Itanium to Intel?
Incredibly important. It is our desire to move it into the very highest
ends of enterprise computing. You've seen essentially every major
computer company in the world, except one, firmly committed to it. As
you get that kind of commonality, the software environment tends to
follow very rapidly. In this space, it's not just about the
microprocessor; it's about the capability of the platform. That is
jelling now. I think it gets more interesting next year, when you see
second-generation (McKinley) systems, especially coupled with
technologies such as Infiniband. It becomes very interesting in terms of
the ability to run every aspect of computing in this architecture.
How did you arrive at the recent deal with Compaq, which plans to
phase out its Alpha chip in favor of Itanium?
Suffice it to say that doing proprietary microprocessors is a bloody
expensive business. Doing any microprocessor is a bloody expensive
business. If you're amortizing that cost over a few units, it's even
more prohibitive. Differentiation is not in the processor. Compaq came
to the same conclusion that HP did--which is that I can use my
engineers to do other stuff because Intel can do (the chips) better and
cheaper.
Where is Intel going with its mobile processors? My sense is that
Intel will begin segmenting that market even more, starting in 2003 with
the successor to the Pentium III-M. Will we see a new Pentium brand at
that point?
It's too soon to predict that. There will be no brand before there is
time. I do think the trend is correct in that we're now developing
specific products for mobile. Ultimately, the ability to build in
seamless connectivity and to have very low power consumption are the
vectors that we have the team on today.
Realizing that decisions on adopting technology must be made years
ahead, we'll just cut to the chase on this one. Was Rambus a mistake?
Hindsight is wonderful. The technology is not problematic. The
technology is still the best technology out there for showing off the
microprocessor. The problem is the cost structure. The cost structure
never met the goals we had.
I look back and say for the time, I really think it was the right
technology.
What's the future for the Architecture Group?
Some analysts have
suggested it would become a cash cow of sorts for Intel, meaning that it
will provide steady revenues for the company but grow at a much slower
pace than other businesses. Our goal is to grow as quickly as the computing markets we serve will allow and to grow our revenue per computer over time. I absolutely do not view Intel Architecture Group as a cash cow. We are investing to grow our business and capture new revenues.
Following up, has Intel's overall Internet building blocks approach
been sidetracked by the lousy semiconductor market of late? Any changes to
that strategy in the works?
There have been no changes to our fundamental strategies.
What's next, Paul? Are you Craig Barrett's successor?
You'll have to ask Craig and the board. I don't think I'll get a vote.
Craig has said very skillfully and elegantly that there are a number of
senior managers at Intel that have the potential to move up.