"On the wisdom with which we bring science to bear against the problems of the coming years depends in large measure our future as a nation."
With that statement, Vannevar Bush, President Franklin D. Roosevelt's science adviser, concluded a seminal 1945 report in which he made the national security case for supporting the kinds of contributions that science had made to the winning of World War II. A major effort in R&D, Bush asserted, would be needed if the U.S. were to continue to prosper.
Some U.S. leaders make the same case today.
Bush's report led to the creation of the National Science Foundation (NSF) in 1950 to support and coordinate education and fundamental research in science and engineering. The belief was that new knowledge and new research would strengthen the nation both militarily and economically.
For validation of that R&D premise, look no further than the post-World War II successes of the U.S. technology-driven manufacturing economy. Bar codes, CAD-CAM, fiber optics, the Internet, Web browsers and many other technologies key to the strength of U.S. manufacturing were born of the public-private partnerships coordinated by the NSF as well as by other government-sponsored research initiatives. And if you missed the recent headlines, technology was the decisive factor in the U.S. war against Iraq and terrorist cells in Afghanistan.
However, many leaders in industry and government fear that current R&D practices will hinder the country's ability to repeat such commercial and military successes. Being questioned is whether the U.S. can keep hold of technological preeminence while production increasingly is conducted off shore, or whether innovation will inevitably follow production beyond U.S. borders. Consider this headlined warning: "DOD Technology Advisory Group Says Military Capability Is In Doubt Due To Loss Of Electronics Industry" (from "Manufacturing & Technology News," May 16, 2003). Drawn from a "controversial and unpublished" briefing document, the newsletter's story quotes a report from the Pentagon's Advisory Group on Electron Devices (AGED): "Off-shore movement of intellectual capital and industrial capability, particularly in microelectronics, has impacted the ability of the U.S. to research and produce the best technologies and products for the nation and the warfighter." The report further contends: Eroding technological leadership in electronics and semiconductors is forcing the Defense Department to obtain the most advanced technology from overseas.
AGED's warning is not the first to highlight the national and economic security concerns arising from a decline of U.S. manufacturing in general and the semiconductor industry specifically; nor is it the only one currently making the rounds in government and industry. For years, concerns that technology and innovation would follow production to low-cost developing countries have been building, as first Japan then Taiwan, South Korea and the other "Asian-tiger" economies developed formidable high-tech manufacturing sectors. Throughout these challenges, however, U.S. producers retained their advanced manufacturing leadership and, in fact, became stronger by taking advantage of the new technological abilities of emerging economies. Consensus has been that U.S. manufacturers would outsource low-skilled, low-paying production jobs while retaining the high-paying engineering and design (as well as finance, sales and marketing) positions. Also, as technology aged. it would flow to less-developed countries, freeing U.S. minds to pioneer the next big market. Those who dared to raise concerns that such a strategy might endanger U.S. manufacturing were written off as protectionist and isolationist.
But the fast entry of China into manufacturing, and especially into the semiconductor industry -- and China's aggressive strategy to encourage manufacturing growth -- has renewed the debate: Are U.S. businesses and the government doing enough to retain leadership in manufacturing innovation?
Reports from the National Association of Manufacturers (NAM) and the National Coalition for Advanced Manufacturing (NACFAM), both in Washington, D.C., suggest earlier fears were well-founded. U.S. leadership in several critical industries is falling along with market share, and high-skilled jobs and innovative research are now following production jobs. Further, the reports highlight that stakes are higher this time, with the entrance of China. The NACFAM report, for example, notes that while Japanese, Taiwanese and South Korean wages rose as those countries developed, the sheer size of China's population means its labor surplus, and thus low wages, will put pressure on U.S. manufacturers for a long time to come.
But it's the potential loss of the semiconductor industry, an industry critical to maintaining America's manufacturing and military strength, that has become the rallying point for those concerned about U.S. manufacturing.
In fact, for 15 years, some people have been sounding alarms about the future of the U.S. semiconductor industry. For example, in 1989, Robert B. Reich, then a professor at Harvard University's Kennedy School of Government, warned that the U.S. semiconductor industry was losing its competitiveness in the world market. In a "Scientific American" article, he stated that between 1984 and 1988, the U.S. share had fallen to 37% from 50%, while Japan's share had risen to more than 45%. Things have become markedly worse since Reich's warning. This past May a National Academy of Science (NAS) report estimated U.S. share of global semiconductor production at 20% to 25% -- and contended the decline was continuing.
Two basic challenges confront U.S. semiconductor producers, says industry trade expert Thomas Howell, a partner at Dewey Ballantine LLP, Washington, D.C., and a contributor to the NAS report on the semiconductor industry.
"The first is the rapidly rising cost of manufacturing semiconductors as feature size shrinks. Line width is becoming smaller and smaller and is now reaching metrics that are nearing molecular and atomic levels. That constitutes a technological 'brick wall' for silicon technologies. Adding functions can no longer be done through the conventional processes of miniaturization." Unfortunately R&D on a scale necessary to find a replacement is not being funded in the U.S., he says.
"The second challenge is that the actual cost of producing these devices keeps growing exponentially," says Howell. "The cost of a state-of-the-art fab is now several billion dollars, [an amount] many companies are reluctant to spend -- especially since the generation life of a device is only a few years."
Another challenge is the semiconductor sector's declining revenue growth rate, asserts Randy Isaac, vice president of strategic alliances for IBM Corp.'s Technology Group, Somers, N.Y. "The rate has dropped from double-digit ranges to single digits in the last few years while the costs of manufacturing and the cost of development are growing rapidly."
Howell says the challenges have led to the semiconductor business model that disconnects device design from manufacturing. "Semiconductor facilities are opening as contract manufacturers -- often with a [foreign] government's backing. "With this business model, first seen in Taiwan, semiconductor designers can outsource manufacturing and avoid investments in semiconductor equipment."
Howell estimates that Taiwan has already captured two-thirds to three-quarters of the world's entire foundry work. He notes that China is rapidly emulating that success with expertise from Taiwan, with the development of special government-funded industrial parks, the low costs of building construction in China as compared with the U.S., and less onerous environmental regulation. Says Howell, "Expect China to be a major semiconductor competitor [of the U.S.] within 10 years."
IBM's Isaac wonders if the U.S. survivors can successfully concentrate on R&D and outsource manufacturing. He observes that the coupling between R&D and manufacturing is growing tighter as the technology progresses. "Feedback from manufacturing experience is vital to R&D, while new innovation needs to be implemented in manufacturing very quickly."
Howell says the question everyone is asking is how long will it take for these contract foundries to also capture the semiconductor R&D. He also sees the move to off-shore contract foundries cascading into a powerful draw for engineering and research professionals. Already multinational companies, such as General Electric Co. and Microsoft Corp., employ scientists and engineers in research labs in several foreign countries, including low-cost China and India. According to one report, over a third of Microsoft's 180 programmers in its Beijing research lab have Ph.D.s from U.S. universities. Indeed, China is aggressively pursuing such individuals by offering incentives such as tax benefits, world-class housing and extensive stock options.
At the very least, these developments will substantially reshape the global semiconductor industry. Isaac sees the U.S. industry responding by aggressively seeking lower-cost manufacturing, more efficient manufacturing processes and by creating alliances with other companies. "Some countries offer attractive incentives to encourage semiconductor manufacturing to be located in their geography," adds Isaac. The trouble with that is, according to a National Institute of Standards and Technology report: "If domestic R&D resources are not available, U.S. companies do not hesitate to form research partnerships with foreign companies, outsource R&D overseas or directly invest in foreign research facilities. These research relationships often lead to follow-on foreign manufacturing relationships. Thus, the maintenance of an effective domestic R&D network is essential for attracting domestic and foreign R&D funds and subsequent manufacturing."
Sen. Joseph Lieberman (D, Conn.), drawing upon a letter from the Semiconductor Industry Association to U.S. Trade Representative Robert Zoellick late last year, points out that China is using "direct and indirect subsidies" to attract next-generation semiconductor fabs, including offering "a substantial rebate on the value-added tax (VAT) charged on Chinese-made chips."
The question of whether or not semiconductor leadership will migrate to other parts of the world has spawned several efforts by governmental organizations to research the issue on behalf of the semiconductor as well as other high-tech manufacturing industries.
Another danger of relying on off-shore manufacturing is the possibility that other countries will capture tech know-how by taking it from the U.S. firms' production facility. It's no secret that China, like other developing countries before it, allows foreign joint ventures into the country in part to gain technology transfer. It's also generally agreed that the country's enforcement of intellectual property protection is woefully inadequate, though it has made progress, enacting better laws and pledging better enforcement. Still, others wonder if China's trade officials aren't stalling and turning a blind eye toward offenders.
The dilemma that results from separating manufacturing from R&D is not confined to the semiconductor sector. It reaches across the full range of U.S. industry, notes economist Joel Popkin in his new study for NAM. His research "Securing America's Future: The Case for a Strong Manufacturing Base" identifies manufacturing as a significant key to innovation at all levels.
Popkin states that "manufacturers' investment in physical and human capital, R&D and productivity are intertwined and together provide substantial economic benefits. This is the method by which innovations become an integral part of the economic process and lead to widespread improvements in productivity."
Actually the benefits are far greater than what productivity might imply, says John A. McFarland, president and CEO, Baldor Electric Co., Fort Smith, Ark. McFarland, whose mantra is "I don't want to sell Chinese motors to my customers," passionately believes that separating research and product development from manufacturing does customers an injustice in terms of product value and service. (He claims the shortest lead times in the industry.)
His argument: "When you go to China, India or wherever, why are you going? You're going to achieve one thing -- a better cost. However the customer does not make his decision to buy based strictly on cost. Customers are also interested in other product attributes. These include quality, performance, availability plus a lot of other issues that don't get resolved by outsourcing manufacturing."
McFarland maintains that the lower pricing might be one -- but not all -- of the significant factors in any buying decision. "We believe that we can satisfy the other three-fourths by producing the products where customers, research and development, and manufacturing can easily and quickly interact. Being price competitive can be resolved with R&D, product development, employee training and investments in automation equipment." He says the greatest product cost improvements at Baldor have been made through R&D improving the product design.
McFarland sees time and product variety as two emerging issues in the marketing of industrial products. He asks: "How do you offer more product variety and simultaneously improve the time issue by moving the manufacturing site thousands of miles from the R&D function? At Baldor the R&D function sits right next door to the plant. We can design a new motor in the morning, test it in the afternoon and have it in production the following day. With globalization, it is very tempting to think that long production runs of a small number of products will satisfy the market. That's not the way our customers buy."
Popkin's research emphasizes that manufacturers invent the future by closely integrating R&D with manufacturing. He says manufacturers are responsible for almost two-thirds of all private sector R&D -- $127 billion in 2002. Spillovers from this R&D benefit other manufacturers and non-manufacturing firms because, Popkin explains, they are magnified by geographic proximity, enabling innovation in one area to freely stimulate products and processes in other areas.
Popkin's study warns that if the U.S. manufacturing base continues to shrink at its present rate, and the critical mass is lost, the manufacturing innovation process will shift to other global centers.
Steve Goldman views his new fabless semiconductor business model as passionately as Baldor's McFarland views servicing his customers with U.S.-made products. Both see faster economic gains with their approaches. The difference is that Goldman doesn't share McFarland's insistence on keeping R&D and manufacturing united in the U.S.
Chairman and CEO of Power-One Inc., Camarillo, Calif., Goldman is convinced that by going fabless for a new product line, he is gaining value for all his corporate constituencies.
His company, one of the 10 largest power conversion equipment vendors, is ranked in the S&P 500. The company's whole focus is on powering the communications and technology infrastructure -- everything from small networking equipment to cell stations and optical backbone. It has sales, manufacturing and R&D spread across China, the Dominican Republic, Ireland, Norway, Slovakia, Switzerland and the U.S.
Goldman says his company's maXyz (pronounced max-sys) product, on which patents are pending, differentiates Power-One by placing it in the silicon-based power conversion and management market. The device, the size of a contact lens, is a power unit for telecommunications and replaces conventional circuit-board-mounted electronic components. While maXyz is being outsourced, the product nonetheless represents the biggest R&D initiative in the company's 30-year history. The R&D investment represents almost 17% of sales in an industry where 4%-5% is traditional.
Goldman describes the maXyz R&D initiative as a strategy for uniquely repositioning the company for accelerated success as the economy recovers. "Our mission was to use our R&D initiative to change the [competitive] playing field." As a result of bringing in new technology Goldman says all of his traditional competitors are no longer a business factor for the maXyz line. "Our new competitors are semiconductor companies."
He fervently believes that the R&D initiative is the fundamental component for Power-One's business strategy. "We began our investment at the right time, with the right technology, looking not too far ahead and not too close."
To make sure technology is the answer meant Power-One had to use maXyz to depart from the industry norm of being an assembler and integrator of the intellectual property of other companies. "With maXyz we will no longer buy silicon chips and components from big corporations and integrate them on a circuit board to provide a total solution for the customer."
Why change?
"The profit margin wasn't there. The people making the real money are those supplying those components, who were developing intellectual property (IP) and patents with 45%-60% gross margins. We were achieving 40% gross margins at best, but now with our own IP we can raise that to 50% plus. The industry norm is about 25%. So instead of being an IP aggregator, we're evolving as an IP generator. In innovating maXyz we've developed more patents in the last 12 months than in the last 25 years."
The differing approaches of Goldman and McFarland raise three critical questions for U.S. manufacturers:
How U.S. manufacturing executives answer those questions will shape America's future as a nation. Vannevar Bush's call for leadership in science is as important today as it was more than a half century ago.