It's no longer U.S. vs. them in chip industry
Continuing advances are now beyond the grasp of Americans alone
BY JOHN MARKOFF
New York Times
A decade ago, when a Japanese company sought to acquire a Silicon Valley operation making crucial equipment for semiconductor manufacturing, the prospect was treated in this country as a threat to national security.
Chip industry executives and politicians feared that the sale would touch off a chain of events ending an era of American technological supremacy.
Now those fears have all but vanished. And a growing number of industry executives say the most important lesson is not the threat of overseas competition, but the value of international cooperation.
The reason is a practical one: It will require almost $1 billion to develop the essential technology for the next generation of chip-making equipment. And with fears of Japanese domination having abated, the U.S. government is underwriting less of the underlying research. So chip makers say continuing advances are now beyond the grasp of Americans alone.
Indeed, an alliance of American, Dutch and German chip makers is expected to reach a milestone in April when they switch on a futuristic chip-making machine able to etch circuit lines no more than several hundred atoms across in width. It is a commercial engineering challenge -- and an international collaboration -- on a scale rare if not unequaled in any other industry.
``The semiconductor industry has been unique in pursuing this form of global cooperation,'' said Ulrich Schumacher, president and chief executive of the German participant, Infineon Technologies. ``It's because our costs are brutally high.''
The privately financed cooperative effort is being conducted in national laboratories better known for their work on the American nuclear arsenal. It harnesses optical technologies involving highly reflective mirrors and high-powered laser light sources borrowed from research done as part of the missile-defense program known as Star Wars.
The resulting chip-fabrication plants -- which themselves will cost $1 billion to $2 billion each -- will be able to produce chips for a myriad of consumer applications, from PCs to cell phones, that will be 100 times as fast as those in use today, with 100 times the storage capacity.
While the burden of developing chip technologies has encouraged a trend toward global cooperation, it has also led to the emergence of new players in the race to build state-of-the-art computer chips. In the last decade the semiconductor industry has expanded dramatically with aggressive new chip makers emerging in Taiwan, Korea and Europe.
The globalization of the chip industry is accelerating in part because other regions are adopting gadgets like cell phones and wireless digital devices at a faster rate than the United States. But the role of American chip makers in sharing and transferring technology has also been a factor. IBM, for example, rather than bear the cost of research and development on memory chips by itself, entered into a joint venture with Infineon, and the partnership was instrumental in helping make Infineon a world leader in memory technology.
And though the result has indeed been a dilution of American leadership in the chip industry, there has been almost no evidence of the anxiety that shaped a national debate on industrial policy in the United States a decade ago.
``There was a concern in Washington that this had to be a U.S.-only technology, and we've been working hard to explain to them that this has to be an international technology,'' said Charles Gwyn, program director of the American-European consortium working on the next generation of chip making gear. ``It needs to be accepted worldwide to be successful.''
In 1999 global sales of the semiconductor equipment industry were $28.6 billion and worldwide chip sales reached $149 billion. Growth rates for chip production and chip-equipment purchases in Europe, Korea and Taiwan indicate that those regions are rapidly catching Japan and the United States, particularly in the most advanced memory chips -- widely used in the most popular consumer products, from desktop PCs to digital music players -- and in state-of-the-art factories.
Today three European chip makers -- Infineon, ST Microelectronics and Philips -- are among the world's 10 leading chip makers. Meanwhile, a decade after the Japanese had come to dominate the memory chip sector and were threatening to overtake the United States in microprocessors, they are now being outpaced, in part because they have tried to go it alone rather than cooperate in research and development, Schumacher said.
``They thought it was about controlling the technology,'' he said.
The shift toward collaboration can be seen most clearly here in the San Francisco Bay Area, where the privately financed international effort to design the next generation of chip making gear is under way at three national laboratories, Lawrence Livermore, Lawrence Berkeley and Sandia California.
Infineon of Germany, ASM Lithography of the Netherlands and three American chip makers -- Intel, Advanced Micro Devices and Motorola -- are underwriting the initiative, which involves an exotic new technology known as extreme ultraviolet lithography, or EUV. The research is considered crucial to continuing advances in the semiconductor industry beyond 2004, when current technology is expected to reach its limits in etching ever-smaller circuits on silicon.
When Robert Noyce, co-inventor of the silicon chip, set out four decades ago to build one of the first chip-making machines, he drove to a photographic supply house in search of an inexpensive used camera lens for his new machine.
In contrast, the new machines, called steppers, are custom-made devices. They are based on a laser-generated light source at wavelengths far shorter than those commonly used today. And they require engineering feats, like the ability to create a machine that can send light through a vacuum -- and one that moves fast but produces no vibration.
The consortium researchers acknowledge that they still have thorny technical problems to solve. And IBM and Lucent Technologies are pursuing an alternative direction, using ultra-short-wavelength X-rays to etch the circuits, instead of light.
But the Semiconductor Industry Association points to EUV as the more promising technology, and the confidence of the California-based researchers has been growing in recent months. Their approach is expected to lead to a prototype machine by April. If all goes well, such machines will be producing chips commercially within five years.
The company that figures to bring the technology to market is, in fact, the Silicon Valley Group, the outfit that caused a furor when it was coveted by Nikon of Japan a decade ago. At the time it was the lithographic division of Perkin Elmer; IBM ultimately stepped in to ensure that it stayed in American hands. But after further changes in ownership, it was recently sold to ASM.
At the same time, Japan is independently pursuing the EUV technology in a government-financed consortium. Whichever team gets there first, it appears likely that for the first time there will not be an independent American-owned maker of the most advanced chip-making machines.
Today the new landscape is being hailed by many in the computer industry as a major step toward a borderless world in which economic and technical interdependence are the hallmark of a new global economy.
``I really think the world has changed for the better,'' said Papkin Der Torosian, the chief executive of Silicon Valley Group.
But some American officials remain concerned that while government-financed consortiums in Japan and Europe are continuing to spend hundreds of millions of dollars on new technology, U.S. government funds for research in semiconductor manufacturing are dwindling.
In the early 1990s, the Defense Advanced Research Projects Agency was spending more than $300 million a year on efforts to develop new semiconductor capabilities, in part to ensure that the Pentagon would not have to turn to foreign suppliers for its own needs. That spending is now $55 million a year, and is scheduled to be phased out by 2005.
``The United States programs in basic research are a lot smaller than the Japanese or European efforts,'' said Gordon Moore, a co-founder of Intel. ``But people here are riding high and so there just isn't a great deal of concern right now.''
There is another, more fundamental reason there has been no political outcry about the new globalization of the chip field, according to many industry executives.
The rapid internationalization of the industry has made it possible for large and small American companies to take advantage of the low and subsidized cost of capital in countries like Taiwan and Korea.
``The industry has become so integrated globally that what was a major trade issue a decade ago is today scarcely a peep or a murmur,'' said Kenneth Flamm, a former Pentagon official who is now a professor of international affairs at the University of Texas at Austin. |
