Imagine life today if the integrated circuit had never been invented: no personal computers or Internet,no cell phones, no satellite weather maps, no Mars rovers. The IC has generated applications that were never imagined circa 1959, when Jack Kilby at Texas Instruments and Robert Noyce at Fairchild Semiconductor separately filed for patents on it. The basic invention took off on its own and created a highly unpredictable race course for technology in general.
A fundamental shift in spending patterns by private industry and U.S. agencies now threatens to put the brakes on just that kind of radical technology advance. There is a growing consensus among academics, CEOs and industry lab directors that the United States is not adequately funding basic research, while instituting policies that, perhaps inadvertently, are strangling the culture of innovation here.
"We had a sensitive but stable ecology: Government gets taxpayer money, invests it in long-term research; some of that leads to these industries' getting started, which generates more tax money. This is the right model," said David Patterson, a professor of computer science at the University of California at Berkeley. But government planners appear to be backing away from that model, and Patterson-who has become an outspoken critic of what he sees as a dangerous shift in U.S. government spending on basic research-is not sure why.
"For some reason, there is this contrarian view that says, 'Hey, look at all these highly competitive industries with very low margins-they have a lot of money, they should be paying for this,' " Patterson said. "But it doesn't work. There is no way these companies with continually shrinking margins are going to fund fundamental research."
A recent study by the Government Electronics & Information Technology Industries Association (Arlington, Va.) predicts "sharp reductions" in Department of Defense R&D spending by the end of the decade (see www.eet.com/news/latest/technology/ showArticle.jhtml?articleID=172900478). The DOD accounts for the lion's share of federal spending on research. The GEIA report cites the budget deficit and the war in Iraq as among the causes. Also, a new policy that seeks to get equipment in the field as soon as possible is eating up funds that might otherwise have gone to R&D.
Moreover, the government has become much more aggressive in classifying projects as top secret, discouraging academic participation, restricting participation by foreigners and instituting a new policy that subjects research projects to a go/no-go review after only 12 months, experts in industry research said.
Noting these trends, the Semiconductor Industry Association recently launched a blue-ribbon committee and lobbying initiative to sound the alarm about gaps in basic research (see Sept. 19, page 1).
It's not just the dollar amount allocated to research that has UC Berkeley's Patterson worried. He is also concerned about what appears to be a seismic shift in government spending patterns.
Two main funding agencies, the National Science Foundation and the Defense Advanced Research Projects Agency (Darpa), have taken complementary approaches to basic research. The NSF follows the mainly academic route of subjecting proposals to peer review. That agency continues to fund basic research at about the same level as before, but NSF covers a much wider field than the engineering-focused research at Darpa.
Darpa, for its part, has long taken the brasher corporate model of backing visionary projects created by bold thinkers. The Internet, graphical user interfaces, the single-user workstation and the entire field of computational science are some of its success stories. But more recently, Darpa has backed off from its bold approach, according to Patterson and the heads of a number of industry labs around the country.
"Darpa has really shifted its emphasis, and we don't really understand the agenda," said Glenn Edens, director of Sun Microsystems Laboratories (Menlo Park, Calif.). "But if you talk to a lot of the local universities in Northern California, they would tell you that a lot of the Darpa funding that has been really important to computer science and the Internet has dried up."
"For the semiconductor industry in general, but also here at TI, we feel that the investment in long-term engineering fields has not even kept up with GDP [gross domestic product] kinds of levels," said Hans Stork, chief technical officer at Texas Instruments Inc.
In addition, observers said, a crisis may be developing in terms of the talent pool. The United States has traditionally attracted top talent from around the world and developed unique, concentrated cultures such as Silicon Valley, where fundamental ideas incubate. The erosion of research here may accelerate as talented people find better opportunities elsewhere.
"There are still companies doing applied research in the U.S., but I think the biggest significant difference [today] is that some of the developing countries that are in economic prosperity right now, like China, are spending a good portion of their economic prosperity on applied and basic research," said Artie Chang, CEO of Cradle Technologies Inc. (Sunnyvale, Calif.), a fabless semiconductor company that is working on a next-generation DSP.
Critics say that government policy is also restricting this vital flow of talent, an unintentional outcome of more-restrictive immigration policies resulting from the war on terror. "It's disappointing for us as an industry to see limited access to worldwide talent because of political decisions, basically," said TI's Stork.
"I came to the U.S., and particularly to Silicon Valley, because of the opportunity to associate with the best talent in the world," said Ivo Bolsens, CTO at Xilinx Inc., who has a basic statistical argument for building a culture of innovation rather than simply backing a few people or a single idea. Major breakthroughs that lead to new industries are few and far between, and the chance that a given person or project will produce it is very small. But the greater the number of talented people involved, the higher the chance that something truly significant will emerge, he said.