Chip Design Will Go East, Too

That'll be the next to shift to Asia, says industry vet C.D. Tam

C.D. Tam helped bring semiconductors to Asia. In 1968, the Hong Kong native joined Motorola Inc. just as it was embarking on an early and ambitious Asia-based manufacturing strategy. Three decades later, Tam had risen to the top of Motorola's $10 billion Asian operations. Through the years, he has set up chip-fabrication plants, or fabs, and design centers across the region. One of those is a $1 billion chip fab built in 2001 in the northern Chinese city of Tianjin -- a huge tech bet on China at that time.

These days, China's chip industry is flourishing, and Tam has departed Motorola to run Hong Kong Science & Technology Parks Corp., a new government-backed R&D center. His goal is to turn Hong Kong into a center for semiconductor design. Amid growing optimism that the tech recession is finally over, Tam spoke with Asia Technology Correspondent Bruce Einhorn about the prospects for Asia's chip business.

The recovery that's under way is helping Asian foundries, but what's the impact on Western chipmakers?

It has been a very long recession. But in 2003, all the signs are solid. [With the recovery in gear] it will be a foundry world. Even the mainstream chip manufacturers want to use foundries. Take out Intel, but all the others -- STMicro, AMD, and TI -- are going that way.

How do you regard the rise of China's contract chipmakers Semiconductor Manufacturing International Corp. (SMIC) and Grace Semiconductor Manufacturing Corp.?

In the past, when Motorola built a fab, we built one at a time. But SMIC has four fabs on-site. And Grace is next door with two fabs. There's the Huahong NEC fab, the TSMC [Taiwan Semiconductor Manufacturing Co.] fabs, and now Shenzhen is planning to build three fabs. There are so many things being negotiated right now.

What's the attraction of China?

One thing that has changed [since the downturn] is semiconductor consumption has continued to shift to Asia -- in particular, to the China-Hong Kong area. In 2002, Asia-Pacific excluding Japan accounted for about a third -- $50 billion -- of global chip consumption. In 2005, the region will consume closer to 40% of the world's chips, or $95 billion out of $240 billion total. The consumption will drive the location of the wafer fabs. Chinese chipmakers can now meet about 10% to 15% of domestic consumption. They want to be able to supply 30%. So from a Chinese vantage point, adding capacity makes sense.

How will this shift to China affect other parts of the semiconductor food chain?

This will lead to a major shift of chip-design activities around the world. Most integrated-circuit design has been done in the U.S. and Japan. But in the future most of the consumption will be Asian [excluding Japan]. And half of that will be in China. So chip-design location has to move. How can you design cellular-phone chip sets in Silicon Valley if you only have a chance to visit China twice a year? You have to work very closely with the end-product designer. Which means that the chip-design location will have to get closer to the production sites.

You're promoting Hong Kong as a good place for chip design, thanks to its location in southern China's Pearl River Delta. Yet the city doesn't have a strong manufacturing industry. How will that work?

Seventy percent of China's chip production is consumed in the Pearl River Delta. If a chip-design company sets up in Hong Kong, you're half an hour away across the border from the big customers. Another plus here in Hong Kong is that semiconductor-design companies don't want to have their work copied. Unlike China, Hong Kong has very strong intellectual-property protection.

Some people say the current chip recovery may be short-lived. One of them is Morris Chang, the head of TSMC, Taiwan's largest chipmaker. He sees a slowdown coming as early as 2005, because of China's role. Is he right to be worried?

Governments usually end up going overboard. You just can't tell the Chinese government to slow down the development of the chip industry. So China will create excessive capacity. The next slowdown will come earlier than normal, and China will play a very big part in bringing that about. In two or three years, China will have enough momentum and strong government support. China will overshoot. It won't hurt them that much, but it will hurt others.

    Before it's here, it's on the Bloomberg Terminal.