Raj

Tanna, Director, World Wide Marketing,



(Entry Level Systems), National Semiconductor

With the sub-$1,000 PC market

emerging as the fast-growth track in the PC global market, Cyrix has been dropping CPU

price points at a pace, which has left its competitors struggling. During a recent visit

to India, Raj Tanna elucidated to DATAQUEST, his perspective of the collapsing PC price

structure in the global and domestic markets as well as the technology edge being

exploited by the National-Cyrix merger.

There has been considerable

activity in the lowering of PC price-points in the global and domestic markets. What is

the role Cyrix has played in these developments?



Cyrix has been doing pretty well, starting from last year, when Compaq launched the

Presario in the sub-$1,000 price range, creating that market. There was no sub-$1,000 PC

anywhere in the world before that. The value-proposition that Cyrix brought was the

MediaGX processor, which is an integrated processor. It allowed Compaq to manufacture a

system at $100 less than what an equivalent Socket 7 solution would have costed.

What is the significance of the

Socket 7 architecture, versus the other types, in lowering of price points?



Socket 7 is the standard architecture for the MMX series from Intel, the K5, K6 series
from AMD and M2 series from Cyrix. They have a socket configuration known as the Socket 7

pin out. Motherboards made with processors of this pin-configuration are known as Socket 7

motherboards.

The launch of the sub-$1,000

Presario, forced a lot of our competitors to change their strategies. It forced them to

lower their CPU price points, than what they were normally asking under Socket 7 price

points, to allow them to address the $999 PC market.

When that happened, Compaq took it

a step further, they went to the $799 market with the MediaGX. That forced Intel and AMD

to follow and lower their processor prices to have a $799 Socket 7 system. So our

customers followed with a $599 system. And that's when it really forced Intel in my

opinion, to change their strategy and come out with the Celeron brand to address the

sub-$1,000 PC market.

By then they had seen the data,

they had seen the trends and accepted the fact that the sub-$1,000 PC market is here to

stay.

What is the configuration of

these sub-$1,000 PC's and what has been the uptake from vendors other than Compaq?



If you think about it, last year when Compaq had introduced the Presario, it was 120

Mhz at $999. Today, a $399 PC is running at 200 Mhz, going to 233 Mhz by the end of the

year. Now 233 Mhz for the software that people run at the home today...we are not kidding

anybody, this is not a commercial box. This is a pure consumer box. For people who need to

run word processing, basic internet, some pictures, not some heavy duty database

computing. It runs the basic software that 95% of the people need to run at a very

affordable price.

We have customers in the US who

have introduced systems at $499 and $399 price points, based on the MediaGX processor

. These are full-fledged systems with 32MB RAM, 233 Mhz processor, CD

and floppy drives, speakers, modem, mouse and MS Windows. Everything at $399! That has

taken off in the US and we believe that it is going to take off in India very soon. So you

might expect to see a Rs29,000 PC going to Rs20,000.

Compaq and Packard Bell are heavy

in the consumer market. HP is trying to catch up with the consumer market. These are the

customers, who are talking to us for our products. IBM is heavy in the consumer market and

Acer also.

We have heard about PC

penetration in emerging countries like China and India being highly dependent on price

points. Are you saying that this also applies to developed countries like the US?



Absolutely!

What were they using before?



Nothing! The penetration rate in the US today for the home consumer is still only 43%.

That's a pretty high percentage!





Right, but compare that with 99% for televisions and 75% for white goods. And the number
of homes with a second PC is a very small percentage. We believe in India, we are going to

get to a price point that an average person does not need to think twice about buying a PC

for their child. With schools in India becoming PC literate, it is going to become

difficult for the child to keep up, if there is no PC at home.

How do you convince your

customers about the market size and trend?



According to IDC data, the sub-$1,000 consumer market in the US is 33% of the total market
in 1998. And by the year 2001, it is expected to be 45% of the market. This trend is

followed by some actual data captured by the PC Storeboard. That shows that the sub-$1,000

segment has really grown, the sub-$1,500 has remained flat and sub-$2,000 and sub-$3,000

have really gone down. And this is old data, we are going to sub-$800 and sub-$500 PC.

So how are we going to get to a

sub-$500 PC using a Socket 7 solution. If you think about it, a Socket 7 motherboard today

costs $50, a Socket 7 CPU is about $50, sound and graphics is about $30, that's a $130

solution. For a $599 system if you are going to put a $130 just on the motherboard, add to

that keyboard, mouse, software. It kills you. Its $85 for an OEM copy of Windows 95. Add

to that retail margins, there is no way you can get to a sub-$399-599 PC, using a Socket 7

solution. In MediaGX, where audio is integrated, CPU costs a lot less and the motherboard

a lot less, that's a $80 solution versus a $130 solution. That's what is going to get you

to $399-$599 kind of PC and allow you to make the same kind of margins. l

How have you been able to drop

these price points, in comparison to Intel, considering the amounts they are spending on

their R&D?



That's the problem! They owe something to their stockholders. Stockholders expect a
certain return, which for Intel I believe is greater than 50%. The only way to do that is

by selling processors at a price greater than $200. They absolutely hate the sub-$1,000 PC

market, because this market cannot afford the $200 processor. So they have come out with

this new brand called Celeron. They will take Celeron to the level, which is absolutely

necessary to stay competitive, but no more. They will try to make sure that as much of the

sales as possible can stay at the $200 PII level and as little as possible goes down to

the sub-$1,000 PC market.

How is the relationship between

Cyrix and National Semiconductor being managed. What have you given and what are they

giving to make this merger work?



It's a win-win situation. We bring world class design-we have made plenty of X86
compatible CPU's, the MediaGX system, the integrated proprietary architecture that has

allowed the $399 PC. The whole solution was created by Cyrix. The software, drivers, the

bios, the total system solution. Cyrix did not say, here is a chip Mr HCL, go and create

your bios and your drivers on your own. We gave the whole solution.

So what National saw was leadership

in the sub $1,000 PC-basically X86 expertise. What Cyrix lacked and National brought to

the party was the rest of the pieces. For example, we have an initiative called 'PC on a

chip'. The 'PC on a chip' is going to have all the peripherals on the chip; the super IO,

the chipset, the modem-all the other components that go on a PC besides the processor.

Another factor was, we didn't have a 'fab'. We were dependent on IBM to manufacture the

products for us. That had two disadvantages-cost and time to market. Now with National

having world class manufacturing, we can design our processors for their manufacturing

plants and not any generic fab in the world. Thus they are much more efficient, much more

smaller...

Are you saying the manufacturing

processes were generic because of IBM's plants?



It was generic because we could not design them only for IBM. What if the relationship
between IBM and Cyrix went 'kaput'. Then we would have to go to another 'fab'. And if we

made it a very specific IBM product then we couldn't easily select another 'fab'. We had

to give up on performance and time slice to make it generic. Now we can make it specific

for National. The relationships that National had with tier-1 accounts; Compaq, IBM,

HP-they are big accounts for the other components that National sells-automatically, we

could leverage that it into talking to them about our CPU's.

The biggest of all however, is

Intel cross licensing. Today, National is the only company in the world, other than Intel,

who can make the next generation processor. The slot 1 or the Pentium II, Celeron I or

socket 370, which is the low cost version of Celeron. National is the only company in the

world who is legally allowed to meet that right.

Do you have any plans for Slot1

architecture?



Slot 1 today is not the sub-$1000 PC market. It is not affordable and we don't believe it
is the right architecture for the sub-$1,000 PC market, because of cost factors.

Does National have the capacity

to manufacture sufficient CPUs to meet the global demand levels?



National, when it goes to full capacity, can manufacture 30,000 eight inch wafer charts
per month-I am not saying that we have that capacity as yet. That translates to 3 million

processors per month at full capacity. The annual worldwide demand is about 100 million

processor units. So we are ready to supply these CPU's, keeping in mind that the

sub-$1,000 PC is expected to grow to 50% of the home consumer PC market in the next 5

years. The 100 million we are talking about includes commercial plus consumer; servers,

desktops, portables, servers, everything. The desktop market is expected to be about 80

million and the consumer home market is about 45% of this 80 million. The sub-$1,000 is

about 50% of this consumer market. Now the numbers look very much in reach.

In the past few months there

have been reports of Cyrix processor stock outs. Are you able to meet the demands needs

for the Indian sub-continent?



In the month of August, there was suddenly a total vacuum in Socket 7 when Intel walked
out of MMX and that is why there was a tremendous upsurge. You may be right for two

weeks...it took us some time to get special allocation for India. But since then, it has

been a comfortable position.

The domestic demand for PCs is

850,000 according to IDC (India) for this fiscal year. Entry level out of that is 350,000.

Our goal in the first year of operations in India is to have 25% market share of the entry

level PC market.

Can you comment about the

expected glut in entry-level processors from IBM, IDT and AMD.



IBM had the rights before this to manufacture, market and sell the 686MX processor, which
is really a Cyrix processor. That contract has now been terminated. They are a foundry

partner. They will manufacture processors for us if we so desire, but they will not sell

the product. So that is taken care of. IDT is way behind on their processor road map. We

are at 333Mhz going to 350Mhz, and they are still struggling to make 233 or 240Mhz, which

from our perspective is dead. We don't perceive them to be a threat at this point of time.

AMD is very, very non-prevalent in the Indian market for some reason and I don't know why.

The Indian market has a requirement for a CPU to hit the $50-55 price point for low priced

systems. AMD never goes below $75. Cyrix wants to compete in the $50-80 CPU price points.

With such

a volatile market situation, what is the next six months strategy for India?



The next six months are going to be crucial for us. We are going to be talking to the top
five manufacturers in India, to allow them to make PCs that hit price points that have

never been hit before, with technology that is current. That's is going to allow consumers

to buy affordable PCs. That is going to grow the PC market in India.