Federal Funding Cuts Cause Electronics Brain Drain

Electronic News, April 30, 2001 by Steven Fyffe

Government research funding crisis endangers industry, SIA says

Federal government funding cuts to long-term university research are threatening America's position as the global semiconductor superpower, according to the Semiconductor Industry Association (SIA).

Basic research is needed to avoid the looming physical roadblocks to Moore's Law that could trap the U.S. semiconductor industry in a dead-end as soon as six years down the track. But the federal government is slashing spending on crucial sciences that could provide away out and abandoning private enterprise to shoulder most of the responsibility for research and development, according to a report prepared by three SIA members.

"We are reaching the limits," said Willem P. Roelandts, president and chief executive officer of San Jose-based Xilinx Inc. and the representative for fabless companies on the SIA board of directors. "At some point, we can't shrink any further because of physical, barriers. Unless governments step in and restimulate the long-term research in universities, I believe we are not going to see the rate of innovation we have seen in previous years.

"The government should work through universities to stimulate research. That is an area where the government can really play an important role. The success of the semiconductor industry in this country came because Bell Labs was a regulated company that could afford to do very long-term research."

While many companies sponsor university-based research through organizations such as the SIA-affiliated Semiconductor Research Corp., the research is usually aimed at developing technology that shows almost immediate promise for commercialization. "Any time companies get involved, they are too near-sighted," said Y.C. "Buno" Pati, president and chief executive officer of Numerical Technologies Inc.

Pati has worked on both sides of the fence. Prior to co-founding San Jose-based Numerical Technologies (nasdaq: NMTC), he led a research group at Stanford University and served as assistant professor of electrical engineering and computer science at Harvard University.

"The drive for universities that are trying to get funding is to develop something that is applicable outside academia," Pati said.

During the 1990s, federal funding for math, physics, chemistry and engineering dropped significantly, according to the SIA report. Funding for math declined 20 percent, physics dropped 20 percent, chemistry fell 10 percent and engineering dropped 30 percent to 40 percent, the report says. The San Jose-based SIA is worried the federal funding fall-off could cause severe shortages of skilled workers.

"If you have a decrease in funding for universities, by definition what you're doing is lowering the quality of facilities and professors and basic infrastructure to educate your future engineers or technicians," said Anne Craib, director of international trade and government affairs for the SIA. "This industry has never historically carried the entire burden for doing basic research that may or may never yield any commercial results," she said.

Meanwhile, foreign investment in R&D is on the rise.

"From a negligible patent position in 1982, South Korea has increased its presence in the U.S. patent system by 4,000 percent," according to the report. "They also have at their disposal an educated workforce. Foremost among these newer competitors are the Scandinavian countries, especially Finland, as well as Israel and India."

Faced with the figures, the SIA board of directors passed a resolution to make increasing government funding its primary policy priority, at a meeting in Washington, D.C., in March.

"The SIA board of directors believes that federal funding for fundamental research in physical sciences and engineering performed at America's research universities is the single most important public-policy issue facing the industry today because of the long-term implications for both technology and workforce," the resolution reads.

[Graph omitted]

[Graph omitted]