America's weapons of the future will read "U.S.A." on the outside but they may say "Made in Japan" on the inside. In fact, concerns about the military's growing dependence on foreign electronics prompted a recent study by the Pentagon's Defense Science Board that found that "our capability to field technologically superior weapons may soon be dangerously diminished."
While the report properly cites the dangers of U.S. reliance on foreign military electronics, it incorrectly identifies the Pentagon as the industry's white knight. The report calls for the Pentagon to funnel $2 billion over five years into a joint industry-Defense Department semiconductor research institute and other programs.
The Defense Department's recommended institute closely resembles Sematech, the recently announced industry research and development consortium.
The Pentagon will be the likely channel for any government support of Sematech, placing the military at the heart of industry's revitalization efforts.
The military's lackluster track record in electronics research does little to bolster confidence in the wisdom of putting the Pentagon in charge of an important new development effort. In more than 30 years of generous funding for electronics research, few breakthroughs in semiconductor technology have been directly funded by the Pentagon.
In search of the next leap in electronics during the 1950s, the Pentagon consistently backed the wrong horse.
Neither the Army's "micromodule" program nor the Air Force's "molecular electronics" research, which were efforts to miniaturize electronic components, paid dividends. Instead, Texas Instruments patented the integrated circuit in 1959, initially without government support.
Similarly, other major innovations such as the planar process for mass production of silicon chips, metal oxide semiconductors and the microprocessor were corporate--not military--projects.
The Pentagon did play a positive role in nurturing the infant semiconductor industry, buying 70% of the early chips. This support allowed the industry to climb the learning curve and gain a sizable edge in world markets.
But military and commercial markets have since diverged sharply, leaving the Pentagon with little positive influence over emerging commercial technologies.
The recommended massive new Pentagon research program may sire a new generation of military chips but is likely to have little effect on the commercial industry.
Pentagon research channels technology toward baroque military applications, which are seldom commercially useful.
Military chips are required to perform highly specialized, rigorous tasks and are therefore designed with customized, rigid hardware. In contrast, successful commercial chips typically have flexible hardware that can be adapted with software for specific applications.
In addition, the Pentagon requires semiconductors that have some markedly uncommercial characteristics, such as designs to guard against radiation in the event of a nuclear war. Also, military chips must meet more rigorous specifications than their commercial counterparts.
For example, many military chips must function from 55 degrees below zero to 125 degrees above, a spread far greater than necessary for commercial chips.
The Pentagon wants new chips that are faster and more reliable, not necessarily less expensive. U.S. industry at present lags not so much in developing innovative products as in manufacturing them cheaply. Because Defense Department officials are relatively insensitive to costs, they can't guide U.S. industry toward more efficient manufacturing.
Military research is often shrouded in secrecy, so innovations trickle out slowly to the private sector.
Pentagon research also tends to favor well-entrenched companies at the expense of the small, entrepreneurial firms most likely to make important new discoveries.
The military's current leading effort to develop a new generation of speedy military chips is the Very High Speed Integrated Circuit program. In 1980, the Defense Department's under secretary for research and engineering, William J. Perry, promised that this program would "ensure that the U.S. maintain a commanding lead in semiconductor technology."
Far from achieving that goal, VHSIC appears to confirm the dangers of the Pentagon acting as a high-tech patron.
The VHSIC program is behind schedule and is projected to cost nearly $1 billion by 1990, triple its original estimate. Of the six VHSIC contractors, five have produced chips specialized for military uses, while only one has produced a chip flexible enough for widespread commercial use. Given the importance of the semiconductor industry to our economy and national defense, it is appropriate for the government to play an active role in revitalizing the industry.
Re-establishing U.S. electronics leadership will require close cooperation between industry and government on a wide range of issues, from trade to tax policy.
Indeed, the military's interests would be best served by the revival of U.S. commercial technology, which could then be adapted to military needs. Giving the Pentagon an even greater role in picking technological winners and losers in electronics, as recommended by the Defense Science Board, would skew new technology to military uses and retard the recovery of the industry.
Semiconductors, like wars, are too important to leave to the generals.