From tubes to chips

Review of

Making Silicon Valley— Innovation and the Growth of High Tech, 1930-70

Cambridge, MA: MIT Press, 2006, 424 pp.

This is an important book that is likely to be cited in all future work exploring both the origins of the semiconductor industry and the economic and industrial development of the Silicon Valley area in Northern California. Its key contribution is a detailed and rigorously documented examination of the linkages between the infrastructure of the vacuum tube industry that emerged in Northern California over 1920-1950 and the later development of a semiconductor industry in Silicon Valley. Chapters in the book cover the pre-1970s history of Litton Industries, Varian Associates, and Fairchild Semiconductor, with occasional digressions on the early history of semiconductor firms Amelco/Teledyne, Signetics, Intersil, National Semiconductor, and Intel.

What this book does best is to show how the 1920s San Francisco Bay Area radio manufacturing and amateur radio communities provided the foundation for the wave of innovative startup companies involved in high-performance vacuum tube manufacturing. These Bay Area companies (Eitel-McCullough, Litton Engineering, ICE, Heintz and Kaufman, and Federal Telegraph) grew greatly during World War II, propelled by a continuous flow of military procurement contracts, and the invention and mass production of radars, jammers, and communications gear generating the demand for their tubes. Other electronics or radio firms— including Sperry Gyroscope, Philco, and ITT—had purchased local players or set up Bay Area operations to tap into the region’s growing knowledge base before the war. Stanford University also played an important supporting role in these developments.

After the war, with a transition to a peacetime economy, only the companies that continued to innovate and interest military users in new tubes for military applications—notably Litton, Eitel-McCullough, and Varian Associates (whose founders left Sperry Gyroscope)—prospered. With the Korean War and the Cold War, business was soon booming again. East Coast firms such as Sylvania and GE as well as local start-up Watkins-Johnson established Bay Area operations and participated in the successful pursuit of high-tech military tube markets. Before there ever was a Silicon Valley,there was a “Vacuum Valley,” and Lecuyer shows persuasively how the materials and equipment infrastructure developed for vacuum tube manufacturing became the foundation for semiconductor manufacturing in the 1950s and 1960s.

The book provides the most complete telling to date of the birth of Fairchild Semiconductor, in many respects the poster child for the growth of today’s Silicon Valley in the late 1950s. Lecuyer argues that the restructuring of defense procurement in the McNamara Department of Defense of the early 1960s, with its emphasis on budget cuts, increasing competition, and reducing sole-source contracts, led to huge changes in focus in what was quickly becoming Silicon Valley. Varian Associates merged with Eitel-McCullough and diversified out of defense and into businesses such as TV transmission equipment, vacuum equipment, and scientific and medical instrumentation. Fairchild focused on developing mass-production techniques to slash the cost of its silicon transistors and promoting their use in innovative commercial applications.

Even though the integrated circuit (IC) was invented at Fairchild (and independently by Texas Instruments at the same time), the company was slow to capitalize on it commercially. As a result, a number of Fairchild technical staff left to set up companies such as Amelco/Teledyne, Signetics, and GME. Not until 1964 did Fairchild begin to produce volume IC products, and to cope with its come-from-behind position, it consciously opted to “dump” versions of Signetics designs on the market at half the price charged by Signetics. The strategy was predicated on the idea that Fairchild’s costs would eventually come down below its selling price and that the cut rate prices would greatly expand the commercial market for these products. It was also intended to send a signal about the punishment of defectors and to hurt Signetics financially. All of these things came to pass. The supporting effort to expand commercial markets led to Fairchild’s Gordon Moore writing an article in the trade publication Electronics in 1965, containing a famous prediction of rapidly improving semiconductor performance that became known as Moore’s Law.

In the late 1960s, Moore and others left Fairchild to found Intel, and others left Fairchild to take over East Coast transistor-maker National Semiconductor and make it over into a top Bay Area IC firm. Fueled by stock options, the recurring pattern of innovation, defection, and spinoff of a new startup became the canonical U.S. high-tech growth pattern. Lecuyer ends the book with the argument that the networks of expertise and relationships going back to the vacuum tube business were critical to the development of what became Silicon Valley, that the new silicon infrastructure built on the old vacuum tech base, that taking innovative new products and creating commercial demand for them became a critical ingredient of the new business model, and that the mobility–stock options–spinoff cycle, coupled with an easy flow of know-how and people across corporate boundaries within a region, became central to the new model for high-tech–based economic growth.

The archival research is impeccable, and the book is destined to become an oft-cited resource for those interested in the semiconductor industry. Yet the semiconductor chapters in this book are not without some serious flaws. They omit critical events as well as major players who, although outside the geographical boundaries of Silicon Valley, nonetheless had a major effect on its evolution. Firms such as Texas Instruments, Motorola, IBM, and AT&T Bell Laboratories are referred to only in passing, at best. The actions and policies of these firms in such areas as investments in R&D and licensing of intellectual property had a major role in shaping the development of companies such as Fairchild and the environment in Silicon Valley.

Another omission relates to foreign competitors, who are wholly absent from Lecuyer’s story. This is a serious oversight in an industry that has been, from its start, global in scope. These geographic blinders can sometimes have distorting effects on the narrative. For example, in describing Fairchild’s efforts to go after commercial and consumer transistor markets, the author details Fairchild’s pioneering 1963 decision to establish an offshore plant in Hong Kong in 1963 solely in terms of its impacts on transistor costs. Completely absent from this version of the tale is the fact that Hong Kong by this time was a center for consumer electronics exports to the United States and a natural market for consumer electronics components.

Japanese producers had invested heavily in mastering the production of lower-quality consumer-oriented transistors in the late 1950s and were producing more transistors annually than U.S. firms by 1959. Some 55% were incorporated into transistor radios, of which in excess of 70% were being exported. To ignore the intensifying Japanese competition and the rapid rise of consumer electronics production and component assembly in East Asia at the time when the U.S. semiconductor industry started going offshore is to miss an important piece of the puzzle. Increasing Japanese competence in transistors was also a key reason why the technological transition to integrated circuits was so critical to the future of U.S. semiconductor makers.

Lecuyer’s discussion of Fairchild’s strategy is completely silent on its failed 1962 attempt to enter the Japanese market. Blocked by the Japanese government from investing in a local production facility, Fairchild was maneuvered into transferring the Japanese rights to its patents to NEC for a pittance, on onerous terms that later obstructed its attempts to sell products in Japan.

Lecuyer also wears some disciplinary blinders. Good business history blends the individual and institutional with powerful underlying currents of economic and technological forces. So it is distressing to see the author occasionally struggling to deal with questions to which many economists and economic historians have provided well-developed answers. For example, when asking why it was that Fairchild went into the IC business so slowly and ended up chasing spinoff companies founded by its own defecting talent, Lecuyer cites Harvard innovation guru Clayton Christensen in order to argue that even well-managed companies tend to ignore that which is not needed by their existing customer base. The more compelling economic argument is that the returns from commercializing a new innovation are often lower for an unchallenged industry incumbent as compared with an entrant, because a new product will typically cannibalize the sales (and profits) of older products. A new entrant has no existing sales base and no lost profits to deduct, hence a greater return on introducing the innovation. The likelihood of the incumbent facing a challenge using the new technology is minimal until dissatisfied employees have left, at which point the previous unfelt threat materializes and the incumbent is forced to pull the technology off its shelves and play catch-up. The pattern of defection and spinoff is most definitely not unique to semiconductors and is probably best explained as the predictable result of underlying economic logic rather than an idiosyncratic history of poor decisions.

Some underlying quantitative data on the relative sizes of military R&D and production contracts at various semiconductor firms would have greatly helped with the discussion of the role of military procurement in the development of these firms. At least some of this data is available, and digging it out would have been useful. Given that a small set of firms is the focus of most of the book, even a chronology of major defense programs and their role in technological developments within these companies would have been feasible and useful in advancing the analytical argument.

Those interested in other aspects of the development of Silicon Valley will have to turn elsewhere and make connections to the case studies in this book. The interactions of Stanford University with the development of the Bay Area electrical, electronics, scientific instrument, and materials industries over the course of the entire 20th century are only slightly touched on and could be explored in much greater detail. There is almost nothing about the migration of major defense contractors to the Bay Area during and after World War II.

Notwithstanding my personal list of additional discussion topics and data requests, this is an excellent book. Lecuyer is to be commended.

Cite this Article

Flamm, Kenneth. “From tubes to chips.” Issues in Science and Technology 23, no. 1 (Fall 2006).

Vol. XXIII, No. 1, Fall 2006