In the first crisis, Japanese chip producers raised their share of industry revenues above U.S. producers by the mid-1980s by improving their manufacturing technology, thanks to government demanding technology transfers from IBM wanting access to growing Japanese markets, and lower capital costs. Their government also subsidized R&D, promoted cooperation between competitive groups, and protected Japanese markets. Japanese yields (70 -- 80%) exceeded U.S. (50 -- 60%), and reliability was also higher. American responses included Motorola's Six Sigma program to dramatically improve quality, U.C. Berkeley's establishing best-practice comparative performance benchmarks and the fall of "not-invented-here" attitudes in the industry, lowering the value of the dollar 51% versus the yen, establishing a research consortium (SEMATECH) that helped chip-makers and suppliers work more closely together, and largely exiting the DRAM memory business due to overcapacity.
Intel also ended its "second-sourcing" agreements with other chip-makers, limiting competition, established the principal that the software embedded in the design (microcode) is copyrightable and began challenging threats to its intellectual property, sped up its product development cycle, and began branding its products. Motorola, IBM, and TI waited to exit the market until '97-'98, with only Micron remaining (cut its costs through design). Japan, in turn, lost its DRAM leadership to Korea's Samsung with its greater access to capital, and was further hindered by an overemphasis on quality while the market shifted from long-term mainframe uses to shorter-lived PCs and other consumer goods.
Increasing wafer size and decreasing line-width (from two microns in 1980, to less than .1 micron in 2005 -- human hairs are about 100 microns) led to rapidly rising fabrication plant costs -- the second crisis reviewed. Costs to build a leading-edge fabrication plant rose from $200 million in 1983 to $5-7 billion by 2007. Costs of developing the process flow also rose -- to $2.5 billion. Consolidation was an obvious response, as was the development of independent Asian contract manufacturers that produced chips for other firms. This, in turn, has led to concern over ultimately losing fabrication (and design) skills in the U.S.
The third crisis was the inability to design chips that took advantage of the large number of circuits available through new fabrication plants. This was resolved through system-level integration onto a single chip, thus bringing faster operating speeds, lower power consumption, lower costs, the ability to be used in smaller consumer products, and improved reliability. At the same time, the required software (45% of design cost) rose to over 200 person-years for design and test. This led to the reliance on design automation and reusable code cores. Lowering the cost of design to meet consumer end-product requirements (the market moved from mainly PCs to also include cell-phones, etc.) also led to opening design centers in Asia (especially India), and 24-hour design work through passing the work back and forth between India and the U.S. over the Internet.
Particularly interesting was the authors' pointing out that cost savings via Asian engineering are not as dramatic as they appear -- only 25-50% savings, versus the 80-90% that would appear via salary comparisons. Engineers in China and India, and to a lesser extent Taiwan, are younger with less experience and often hold only two-year degrees. They also often don't get trained in automated chip design. On the other hand, based on Samsung's experience in taking only ten years to move up the DRAM memory technology chain beginning in 1981, the authors envision Asian engineer parity with the U.S. in 10+ years. Meanwhile, the number of U.S. chip engineers has stagnated/fallen, and may already have been overtaken by those in Asia.
A related concern is that outsourcing leading-edge fabrication plants to Asia and Europe will lead to outsourcing chip design as well. The authors sort of go round in circles on this issue, eventually concluding that this negative outcome is likely.
One final challenge: The authors see current consumer-product chips as over-designed -- more functions than most users want. Thus, the next trend will be simplifying those chips for use by poorer populations.
Clair Brown is Professor of Economics and Director, Center for Work, Technology, and Society (CWTS) at the University of California, Berkeley. Her recent research has focused on high-tech workers, firm employment systems and performance, and wage dynamics.
Greg Linden is a Senior Researcher at CWTS and a consultant specializing in the economics of the global electronics industry.
Chips and Change: How Crisis Reshapes the Semiconductor Industry (MIT Press/ Sep 2009) by Clair Brown and Greg Linden
Loyd Eskildson is a regular reviewer of books related to international business and history.
2050: Indian and Chinese Economies Double That of USA
To see more of Basil and Spice, go to http://www.basilandspice.com/ Copyright (c) 2009, Basil and Spice Distributed by McClatchy-Tribune Information Services. For reprints, email tmsreprints@permissionsgroup.com, call 800-374-7985 or 847-635-6550, send a fax to 847-635-6968, or write to The Permissions Group Inc., 1247 Milwaukee Ave., Suite 303, Glenview, IL 60025, USA.

More News:
Market Updates |
Stock Alerts |
All Trading News |
Stock Index