I don't know if anyone has posted this yet....so:|
I pasted all three sections into one long post. Sorry. Kent.
Q&A With Intel's Craig
The last 10 days have been quite eventful for Intel Corp.
First, the company shocked Wall Street by saying
second-quarter sales would be much lower than expected
due to a weakness in Europe. The company's stock took a
big hit, rival Cyrix announced a competing product, and the
company's top brass spent a day in front of hundreds of
financial analysts and money managers. Following the
analysts' meeting in New York last Tuesday, Craig Barrett,
Intel Corp.'s longtim chief operating officer and its newly
appointed president, sat down with EBN editors Jack
Robertson, Ismini Scouras, and Matthew Sheerin to discuss
current business conditions, as well as Intel's strategic plans.
EBN: Why will
second-quarter results be
less than expectations?
What's going on in Europe?
Barrett: What a surprise.
What we do is monitor
billings, bookings, shipments,
and inventory levels, and it just
seemed to have slowed down
in Europe. If I were a stock
analyst, I would say that it is
due to the French election, or
something like that, because it
happened around the same time, and there's a
cause-and-effect relationship. In fact, I don't think anybody
knows. Perhaps we may be back in the traditional European
slowdown. Other than the last couple of years, historically,
Europe has slowed down dramatically in the second quarter.
In the last couple of years, for a variety of reasons, it hasn't. It
may be back to that. Anything else I'd say would just be
speculation. We need some time to study the data.
EBN: How are conditions in North America?
Barrett: North America has not been a big growth site. The
major growth activity has been in the emerging markets,
which is Eastern Europe, Latin American and the Asia Pacific
region. Japan grew very heavy for a couple of years, but now
it slowed down. Fujitsu and NEC price battles made PCs
affordable and it finally made the conversion away from the
Japanese equipment, or the Wang word processors. Those
two things gave an artificial boost to the Japanese PC market,
and it doubled in two years, But the major growth has been in
the India's, Chinas, Koreas and the Taiwans and Brazils and
Russia, and S. Africas - the emerging markets. They're a little
off base, obviously, but their growth rates have been
substantially above the U.S. and Western Europe.
EBN: How fast is the ramp of the speed of
microprocessors moving up?
Barrett: We announced 300 [MHz] with the Pentium II
introduction for availability in the June/July time frame for the
workstation market. Volume shipments of 233 and 266 are
already happening. If you follow Gordon Moore's
[guidelines] they all stay on track, which guides you to the
kind of the doubling of performance and doubling of speed
every two years or so. That gets you up toward a gigahertz
by the end of the decade.
EBN: Is it not a problem to try to keep the memory in synch
with the processor, because the external memory and the
connections are going to slow you down and not be able to
make optimum use of that? You've gone to specify Rambus 2
for later generations of memory.
Barrett: We've said all along that we need to have new
memory-structure capabilities for both memory DRAM and
the backside bus inherent in the P6 architectures, whether it's
a second die on board or whether it's the cartridge we
introduced with the PII, which gives you the opportunity to
speed up or run the SRAM in a simple multiple of the
processor speed, not just off the upside bus to get that
second level cache to run fast.
EBN: Do you buy bare SRAM die to mount in the
Barrett: That will be packaged SRAM die. The Pentium pro
had a dual-cavity ceramic package. The scc cartridge is
basically a little PC board, and you can buy a commercially
available SRAM from a variety of SRAM suppliers to put
that in place.
EBN: Are you trying to narrow down your sources of
SRAMs? Barrett: We will probably will have a limited
number of suppliers. Te don't say how many and we don't
say who they are. We
EBN: Now this is standard SRAM that you're buying from
Barrett: More or less.
EBN: But the memory for Pentium pro was your own
EBN: Was there a reason that you switched back to the
standard with Pentium II?
Barrett: We wanted to use it for cost reasons and the fact
that we didn't want to invest a ton of capacity [on something]
a large number of industry people could produce. We give
out the specs that the part has to be in, and they can bid and
compete with each other to make it.
EBN: What generation do you see the Rambus II or
equivalent memory being in?
Barrett: Essentially you have all of the DRAM suppliers out
looking at what it takes and figuring out what they are going
to do, and Rambus is one solution. There are other solutions
that people are betting on in that area. Everyone is in the
development stages right now.
EBN: You're open, then, to other solutions?
Barrett: We may have a favorite solution, but I think these
things are never over until they're over.
EBN: What time frame?
Barrett: This decade.
EBN: For the first time in several years, you have two
companies competing against some of your product
lines. What's your take on the competition, and how has
that changed your thinking or strategy at Intel?
Barrett: It hasn't changed our strategy at all. Competition is
good for us. One of the ways you keep from getting too
enamored with your success is to have competition. We've
had our product plans and roadmaps laid out for the next
several generations of processors, and the next several
generations of process technology. Our challenge more than
focusing on the competition is to focus on our product lines,
and that's what you're seeing. We introduced the PII. If you
listened in our first quarter conference call, you probably
would have heard that we hadn't even introduced the product
and we shipped a lot more product than AMD, which had
preannounced and hyped the K6. We tend to speak with
products as opposed to paper.
Q&A With Intel's Craig
Barrett - Part II
in light of
Barrett: I don't think so. The way we work is very simply.
We build factories, and we design products and we create
process technology. And our factories crank out product,
and if you look at the marketplace we sell into, there's
specific price points that people sell computers at. If you take
it down the next notch or at specific price points, you can sell
processors into those price points for computers. We tend to
sell out on our factory output, sell the distribution that we can
make, and start at the high-end and [work] down. And you
just scale it from the high-end down to the entry level. So
that's how we adjust our pricing, and that's how we adjust the
product mix and the distribution that we sell. We don't sit
there and hold back 300 MHz. just waiting for AMD to
come out with a new product and say oh, we've got one too.
as soon as it's available, we put it out at the high end and
push everything down.
EBN: Dr. [Andy] Grove mentioned this morning the
company has been committed for years to adhering to
antitrust laws and has several programs in place. Can
you give us some specific examples of some programs?
Barrett: All of our sales personnel have mandatory antitrust
training. Senior managers, including the executive staff, have
mandatory antitrust training each year. One of the things we
do at the executive level is we volunteer one of the executive
staff members each year to a two-hour cross examination by
outside antitrust attorneys on both real and created
documents. This is a blind two-hour interrogation. and
everybody gets to sit around to see how uncomfortable it can
be to try to answer questions if you have done something
We are aware of our market position. We do everything
possible just to keep everybody aware of what's legal, what's
proper and what's appropriate on how to behave yourself.
That system is set up to guarantee people monopolies, and
monopolies are not bad. It's monopolistic power that is
misused, which is bad. So we focus on training people what
they can and can't do.
EBN: And are have there been any businesses that
you've stayed out of because of antitrust concerns?
Barrett: We never even considered not getting into a
business because of antitrust issues. It's a behavior in how
you abuse your position that's important, and we go through
great pains never to abuse our market position.
EBN: You mentioned 300 mm in 1999, 2000?
Barrett: I think that's the time frame I expect to see that in
volume production, probably coincided with the 0.18-micron
stuff. Some people will pick up quarter micron; 0.25-micron
is gone - it's a done deal.
EBN: Are you going to be able to use the SVG
micro-scans on 300 mm?
Barrett: We tend to stick with the suppliers and the road
maps to give us the technology and capability that we need.
EBN: EUV, do you have any partners yet that your
trying to line up in consortium for extreme ultraviolet?
Barrett: The industry is still bouncing around with three
possibilities that exist. There are lots of discussions going on.
We have our favorite, IBM has their favorite, AT&T has their
favorite, and other people will have to pick and chose in
between. We continue to talk to people and coalesce the
industry on a couple of choices. It's our belief that the
industry can't afford to pursue all three of those into a
commercialization venture. The real issue is being able to pick
the right one at the right time and get the industry behind it.
So we have nothing to announce today, we're continuing
discussions with people and are technologists are talking and
we'll see what happens.
EBN: Why has there been a delay with the 440LX chip
Barrett: Is there a delay? We never announce products
before their time.
EBN: But there were expectations that the chip set
would be introduced in the second quarter.
Barrett: People always have expectations of stuff. Our
expectation is to get the LX product line in volume
production by the second half of this year.
EBN: Do you see in your PC video camera any OEM
will try to start bundling or selling that?
Barrett: You've got a bundle today. The people who are
loading the software that Frank [Gill] mentioned are just
throwing them up and saying. "I want to buy an after-market
or an add-on camera." You can get that camera with it today.
EBN: But what took off multimedia was going to
change bundles and all of the sudden people were
getting speakers, CD-ROM's, as they bought their PC.
now the next paradigm is you just add the camera.
Barrett: It's one step at a time. I mean, you get the software
loaded, and then you get some demand out there for it, and
get people to build them. What you need is sufficient volume
in the marketplace and start to get some finite probability of
getting someone who will equip it in their PCs. Then you'll get
enough latent capacity out there, and then you can start
adding cameras at both ends. The model here is very simply
you go to a software bundle step first and then it becomes a
standard after that from a consumer perspective. The cost of
that camera of about a couple of a hundred bucks today is
obviously going to go down in a short period of time, and it's
a relatively trivial deal to add it on.
EBN: When do we reach critical mass?
Barrett: You have to get high-resolution digital cameras
down to an affordable range of a couple of hundred dollars
or less. You need print capability to print out 4X6, 5X 8
photographic quality. Most of these things are graphically
approaching them. If you grab [Kodak chief executive]
George Fisher and ask him what business he's in, he'll tell you
he's not in the paper business, he's in the image business. And
the wave of the future is digital imaging. Everything from
taking them to manipulating them, to saving them, to storing
them to displaying and transporting them. You'll probably
save a lot of trees.
EBN: What do you think DVDs are going to do for the
PC industry and for Intel?
Barrett: It depends how aggressive people are in getting
low-cost DVD players out there attached to PCs. if you can
get that firmly established, and real high-quality digital
resolution stuff on the PC very quickly, it could be big
consumer draw. It would be interesting to see what happens
to digital TV. and how fast that comes in. And what sort of
competition there is between that quality. I mean you've got a
great digital receiver in the PC already.
Q&A With Intel's Craig
Barrett - Part III
EBN: The consumer industry is still
anxiously hanging on to the digital
interlace screen with lower
Barrett: The world is full of religious
EBN: Do you see the possibility of relaying to a screen
in the living room from a PC?
Barrett: There's going to be a lot of activity there. Like
every new product introduction, it's going to depend on how
well it's done and the price points it's at. I don't think there's
enough evidence here, but clearly , a ton of people from the
PC space are going into PC theaters, and a ton of people
from the consumer electronics space are going in the same
direction. It may be a bust, and it may be a great success. But
I think it depends on a bunch of variables that are tough to
EBN: If the computer industry play their cards right,
they have a big opportunity.
Barrett: That's why they're excited about it. But you've got
people coming in at two different directions here. Consumer
electronics guys want modular add-in capability to do this
and each module can't cost more than $500 bucks. The PC
guys want to come in with a $2,000 to $3,000 integrated
EBN: You made an investment in the Samsung fab in
Austin. And I believe a par t of that was to get assured
supply of DRAMs.
Barrett: We're a relatively heavy user of advance memory.
Samsung is the world leader in advanced memory. It seemed
like a logical thing to do.
EBN: You also signed an agreement with them on
future projects working together on consumer
electronics. They also make a competitive Alpha
microprocessor for consumer electronics.
Barrett: Show me two companies that are not competitors,
collaborators, mortal enemies, dearest friends. That's the
nature of the industry. It's incestuous. There's no classic
competitor-vendor-consumer rationale. We compete with
NEC, we supply NEC. Digital and Intel are suing each other,
and we buy Digital stuff for our factory and we sell them
processors and flash and embedded controllers. Nobody
stops and worries about whether they are competing
products that somebody else sells or doesn't sell.
EBN: Do you still see maintaining your market share in
flash? Everybody in the world is getting into it?
Barrett: The DRAM business went bad, and they had to say
there were going to get into something to keep their
shareholders happy. You can remember years ago when the
first time drams went bad, the Japanese were all going to go
into logic stuff, and then the dram got good again, and then
they all forgot about that stuff. and now drams are bad again,
so they are all going to get back into logic or flash. That's the
EBN: Do you see that as hinging on your business
model, where they are trying to put logic with DRAM?
Barrett: There are a lot of smart technical guys that just
don't see how you can be cost effective and combine those
EBN: As far as the cache goes, apparently you feel that
you are using the chips with the bus architectures, and
you don't need to put the SRAM then on the same chip
as the logic.
Barrett: Our strategy is very simple. This year we are
investing more than twice as much as anybody else in this
business. And you need to do that you have to bring up
processor volumes that we've ... You start to add SRAM on
top of that ,and fancy Dram on top of that , and we are
severely limited in the horsepower...TO find people to staff
the factories is that we're building is tough enough. To double
the number of factories by making three of four SRAM chips
for every processor we made as well, I think would be a task
beyond our capabilities. so we're going to use commercially
available stuff as much as possible.
EBN: Does that also go for graphics chips?
Barrett: We've got an announced program in the graphics
area. Lockheed Martin technology relationship. We think
that's an area where from an architecture and capability
standpoint, we can impact the platform and have a
reasonable business as well.
EBN: Are there any carry you through to the next
decade. Do you need any more fabs?
Barrett: We announced the fort worth deal, the
groundbreaking is set up for mid-June, and that site is big
enough for several of these 80,000 sq. foot modules.
EBN: Do you plan to bring up 0.18?
Barrett: It will be a quarter or 0.18, depending on how
demand looks like and available capacity looks like.
EBN: Will there be a fab in Asia?
Barrett: We continue to look around the world. We've got
two international fabs under construction now in Ireland and
Israel. we continue to look. we basically go for the best deal
and the best site we can find.
EBN: Have you picked any fabs for your 300 mm
Barrett: We know where we're going to develop it. But we
can't say. You can probably figure it out. We have
development facilities - one in Santa Clara and one in
Portland. You have a 50-50 chance.
EBN: What challenges do you see on 300 mm?
Barrett: I think its standard stuff. That we have full
equipment sent, and uniformity across the wafer diameter and
all the subtleties that you run. For quite some time, we had a
great 160-mm process running in Ireland, which was not so
attractive when you're using 200-mm wafers to run it. I think
it's usually the things that you don't anticipate which come
back to bite you more than the things you do anticipate,
because things you anticipate you put in a lot of work and
energy. But there is a reasonable cross-industry effort in
terms of equipment and process specifications, and a joint
effort to try to standardize and qualify equipment and that's a
lot different than the either 150-mm or 200-mm stuff we did
in the past. Much of the world will target 0.25-micron
EBN: Are you going to bring it up 0.18? It will be on the
trailing edge of your two-year process.
Barrett: Precisely, precisely.
EBN: How are you enjoying your new position as
Barrett: Oh, it has been fantastic. Since I've taken this
office, we've been sued by Digital, we've had an earnings
shortfall - I can hardly wait for next week! This is baptism by
fire. I don't know how else to describe it.