<<  <  >  >>

Page 17

world. Japanese companies successfully entered that market en masse, causing Silicon

Valley's three largest companies, Intel, Advanced Micro Devices, and National

Semiconductor, to abandon the DRAM market. Intel later acknowledged that it felt it

could have weathered the storm but chose to abandon DRAMs in order to put its full

force behind microprocessor development. What a great decision that was. I was

working in the memory group at Advanced Micro Devices at the time. We did exit the

DRAM business because we could not make money in it. We felt that Japan was dumping

DRAM chips into the United States, selling them below manufacturing cost.

In retrospect, I believe that Japan simply got better at manufacturing than we were

for a while and was able to produce the chips at extremely competitive costs. Charlie

Sporck, then president of National Semiconductor, was the father of Sematech. Sporck

used the U.S. DRAM failure as a rallying cry.

Nishi ran the Toshiba DRAM program, which was the most successful of the

Japanese efforts. He testified that there was very little financial aid from MITI to the

Japanese semiconductor industry and also that the Japanese semiconductor

companies--intense rivals--never shared secret information but only general "road map"

information that allowed the companies to gauge the effectiveness of their programs and

make sure they were headed in the right direction. Three important American

semiconductor companies did remain in the DRAM race: Motorola, Texas Instruments,

and thenstart-up Micron Technology in Boise, Idaho. Texas Instruments now

manufactures DRAMs in plants around the world, and Micron has grown to be a $3 billion

company known to be able to outmanufacture any of its Japanese rivals. The domestic

military chip supply was never in danger, and MITI had very little to do with the Japanese

success in the mid-1980s. Superbly managed Japanese companies simply beat us--for a

while.

The tables have now turned. America again leads Japan in semiconductor market

share. Intel's decision to focus on the microprocessor business, combined with its

excellent execution, has propelled it to become the number-one semiconductor company

in the world. America's semiconductor manufacturing capability has caught up with

Japan's. Our focus on designing innovative chips has proven to be more important than

Japan's focus on grinding out commodity chips at very low cost. Many of the American

semiconductor companies that were very small start-ups at the time of Sematech's

formation--Cypress Semiconductor, Altera, Xilinx, Linear Technology, Maxim, Micron

Technology, LSI Logic, and VLSI Technology--are now substantial corporations with

revenues of from $500 million to $3 billion. Those companies manufacture a dazzling

variety of products. We all export to Japan. The innovativeness and resilience of the

American semiconductor industry enabled it to react to the attack--and win. That success

cannot be attributed to Sematech. None of us were members.

Although the MITI VLSI program was successful, the fact is that MITI has also