Denizens of Silicon Valley spend little time reflecting on its history. The Valley is about anticipating and shaping the future, an activity in which entrepreneurs' pedigrees and even past performance count for little. But for those around the world who want to replicate the region's success, its history is a rich source of information on everything from the roles of government and universities in fostering innovation to developing a venture capital industry and creating clusters of economic activity.
This foreign fascination with America's technology sector launched Christophe Lécuyer on the road to becoming one of the most thoughtful historians of Silicon Valley. Lécuyer was born in 1963 in Paris, where his uncle Jacques Lechalupé was a venture capitalist with Sofinnova Ventures Inc. and occasionally discussed his work with his nephew. Lécuyer attended the Ecole Normale Supérieure, one of France's most prestigious universities, where he studied the history of science. He spent the 1985 academic year as an exchange student at the Massachusetts Institute of Technology, where, at his uncle's suggestion, he explored the relationship between MIT and American industry.
"I was amazed by the scale" of MIT's influence, Lécuyer recalls. "There are so many good schools in France, but their size is about one-tenth of MIT or one-twentieth, so the outcomes are totally different." After Lécuyer graduated from the Ecole Normale Supérieure in 1989, he moved west to Stanford University to study the history of Silicon Valley as a graduate student.
"Once you know the East Coast story, you want to know the West Coast story," he says. "I was interested in Silicon Valley because I'm French, and I was curious to know why things worked out there, and why things were not working in the same way in Europe." Lécuyer received his Ph.D. from Stanford in 1999, and his dissertation served as the basis for two books. MIT Press published the first, "Making Silicon Valley: Innovation and the Growth of High Tech, 1930-1970," in 2005, and last fall issued "Makers of the Microchip: A Documentary History of Fairchild Semiconductor," in which Lécuyer and co-author David Brock edited documents kept by Fairchild Semiconductor Corp. co-founder Jay Last and others and wrote a commentary that doubles as a history of the integrated circuit.
Lécuyer currently works as the principal economic analyst in the Office of the President of the University of California system, and he's started work on a third book, a history of Moore's Law, which Intel Corp. co-founder Gordon Moore first expressed in a 1965 research paper. Moore predicted that the number of components on an integrated circuit -- in other words, computing power -- would double every two years for at least the next decade, a theory that continues to hold true almost a half-century later.
The Deal magazine's David Marcus met Lécuyer at the University of California at Berkeley in July to discuss his research, the enduring fascination with Silicon Valley as a model and the increasing globalization of the technology industry. What follows are excerpts from that conversation.
The Deal magazine: Stanford plays a big role in the mythology of the Valley, as does Frederick Terman, the dean of Stanford's School of Engineering after World War II and the provost of the university between 1955 and 1965. But you have a more complex view.
Christophe Lécuyer: One of the main arguments of my first book is that Silicon Valley didn't come from Stanford. It came from a group of young men who were playing with radio -- the hobbyists -- and that Stanford moved into electronics after these hobbyists. Terman made the connection between Stanford and the hobbyists. He brought in many of the technologies developed in the radio tube industry. He brought in Charles Litton and asked him to build a lab to make vacuum tubes at Stanford in the late 1930s. He did that again with semiconductors in the 1950s. The Terman approach was to learn and bring as much from industry to Stanford as possible.
A large part of your first book and then the second book is devoted to Fairchild. What lessons did you learn from that company?
Fairchild made a name through manufacturing processes. Manufacturing knowledge is central to the rise of Silicon Valley and to the rise of the key firms. Fairchild was critical for two reasons.
They were the first company to create manufacturing processes to make silicon transistors, and then they were the first ones to improve upon that process. They developed the planar process, which is the process that has been used in the semiconductor industry until now. It's clear that Intel will be moving to a new process a few years from now, but the planar process is the process that's been used by everybody over the last 50 years.
In the context of the present debate about the value of patents and innovation, what you're saying is interesting in that you're focusing much more on the Valley's manufacturing expertise as opposed to its development of cutting-edge technology.
Manufacturing technology was cutting-edge technology. In semiconductors into the 1970s, the key technology was proc-ess technology. That's where the innovations were made. Things have changed since then because architecture and design became much more important, but at that time, the process was everything.
At what point does the critical part of what's going on in the Valley shift from manufacturing to more abstract forms of economic activity?
If you look at employment in Silicon Valley by category, manufacturing is by far the largest category until the mid-1980s. From the mid-1980s on, manufacturing starts to decline. There's not much manufacturing left in the Valley. It's an organic process over 25 years, and behind that process there are many phenomena. Other centers of manufacturing grew, especially in Asia. Companies moved things out. The other reason is the shift toward software and industries that are not material and also toward biotech.
And yet manufacturing plays a very important role psychologically in the Valley. People in the Valley still like to think of themselves as making products.
Going back to the 1930s in Silicon Valley, there was a focus on product quality. It was central at Eitel-McCullough or HP or Fairchild. These companies focused on the making of very high-quality products. That was the pride of the people working there. I think this is still the case now. I'm sure that engineers at Google think of their work as making very good products.
Could you discuss your project on the history of Moore's Law?
It's at the very beginning. I think of it in different ways. One is that Moore's Law is a series of statements about the advancement of technology. The statements evolved over time, and Moore himself changed his statement several times. That's one aspect of the story.
Another is the story about manufacturing processes and how to make things smaller and smaller over a very long period of time. It's amazing that this has been going on for 50 years now. People have been able to scale down transistors over that many years, and also put that many transistors on a single piece of silicon. The third part of the story is that Moore's Law became a cultural icon. The idea of the book is to follow these three dimensions of Moore's Law and to bring the story up to the present as much as possible.
For now, [the project is] focused on Silicon Valley, but the idea would be to look at other places where this phenomenon is central. Japan is certainly one of them. The Japanese were really the first ones to use Moore's Law to drive the innovation process. Then there's what happened in Europe, and how people used Moore's Law there.
So the book is partially about the globalization of the technology industry.
You mentioned that you go to Paris, where you grew up, two or three times a year to give speeches on Silicon Valley. So it seems there's still an immense fascination with the Valley as a model.
There is a demand for knowledge about Silicon Valley. Over the last two or three years, the French government, like many European governments, decided that the way to promote economic growth was to promote clusters. There's been a very big effort to promote clusters. There's a demand there for more knowledge about the main cluster, which is Silicon Valley. There are delegations from everywhere to Silicon Valley every day.
What do people take away from these visits to the Valley?
I'm not sure. But it's clear that some people have been able [to learn from Silicon Valley]. To me, it's clear that in Taiwan people have been successful in reproducing aspects of Silicon Valley -- not necessarily the Silicon Valley of today, but the Silicon Valley of the 1960s and 1970s when it was very focused on manufacturing. The key has been a focus on quality and manufacturing and innovation in manufacturing.
The Taiwanese government has been able to bring expertise from Silicon Valley. To me, the key to the success in Taiwan is that they were very good at bringing back Taiwanese engineers who had moved to the U.S., gotten degrees here and then worked for high-tech companies in Silicon Valley and Texas. They were the people who had the knowledge and the contacts. That's essential.
How has research and development evolved in Silicon Valley?
Since the early 1980s, there have been major changes in corporate R&D. Large organizations like IBM and HP and others that were doing a lot of basic research in Silicon Valley have abandoned that field. Now they are focusing on very immediate types of projects. The Silicon Valley of the 1950s and 1960s was based to a large degree on technologies developed at Bell Labs, and many secondary firms recruited people from Bell Labs to run their research labs. But now this source has disappeared, so there's nothing equivalent in the U.S. to a major center of innovation like this that is funded by a corporation and can be used by other companies.
Where does innovation take place now in the U.S.?
It's much more widely dispersed than it used to be. It takes place in lots of companies, including lots of nonmanufacturing companies. It takes place also in companies that sell R&D services, where a lot of innovation is taking place. It's taking place in universities and government labs. It's much more distributed than it used to be, and it's also much less controlled by the federal government. In the 1950s and 1960s, a lot of the work was funded by the federal government.
Is that an aspect of the Valley's development that's been replicated elsewhere? Or is there still an appetite in many countries for a large research institution such as a Bell Labs?
If you look at the French case, innovation is done more in Soviet-like labs. For example, the CEA, a national laboratory devoted to nuclear research, branched out into semiconductors. A lot of the innovation in semiconductors comes from the CEA and the Leti, its microelectronics lab. The Leti is at the origin of many developments in the semiconductor industry in Grenoble. In Germany there are all these institutes, the Fraunhofer institutes, that do research. It's a network of labs that do research for German industry. Each country has its own characteristics.
The ability to attract people from all over the world is one of the main reasons for Silicon Valley's success over the past 20 years. How does that fit into your work about Moore's Law?
One-third of all engineers in Silicon Valley are foreign-born. And if you look at some of the better ones, they come from China or India. It's a major difference from the Silicon Valley of the 1960s and the 1970s, where the great majority of the engineers were born in the U.S., with some people coming from Europe and a few from China. So there's a huge change. It's clear that Silicon Valley companies have been able to bring in all these engineers from places like the Indian Institutes of Technology or from the good Taiwanese universities or Chinese universities, and that's been critical to the rise of many companies. My sense is that there are more and more foreign-born engineers who work in Silicon Valley and then move back to where they come from. They bring with them access to markets, they bring back knowledge, they bring back management expertise, and that creates major competition for Silicon Valley companies.
When did that phenomenon of returning home start in earnest?
There has always been a trickle of people going back. My sense is that movement accelerated after the dot-com bust. Many people lost their jobs. These people had expertise, there was a demand in India or Taiwan for them, so I think this phenomenon of people moving back has grown after 2000 and 2001.
So the bust created a dispersal of expertise and intellectual property.
The dot-com boom was also a boom in engineering salaries, so that the money spent by Silicon Valley companies on salaries grew enormously in this period. There were bidding wars for engineers. It's one of the sources of the offshoring phenomenon after the dot-com bust, where companies try to move research and design groups to India and China as a way of hiring people who are competent at low prices and a way of limiting salary increases in Silicon Valley. Companies didn't want to go back to the dot-com boom period, where they were forced to spend enormous amounts of money on engineers.
There's a long history of offshoring in the Valley. Even in the 1960s, Fairchild and Signetics were moving manufacturing to Asia. Do you see the continuation of this as a real threat to the Valley?
Over the short term, it's a way for companies to maximize their profits. Longer term, I think it's detrimental to the Valley. If you create expertise in other places that are going to compete for people, compete for markets, this is going to have an effect.
How would you contrast the dispersal of expertise we've seen over the past decade with the dispersal of manufacturing expertise that began in the 1960s?
There was a big movement of electronics manufacturing from the U.S. to South Korea, Taiwan, Hong Kong, the Philippines in the 1960s and 1970s. The outcome is where chips are manufactured now. The largest producer of chips in the world is Malaysia. This expertise on how to package the chips, which was moved there in the 1970s by Intel and other firms, remains there. So now Malaysia is the main center of the packaging of chips in the world. I think it's basically the same phenomenon, but what's getting moved now is not the packaging expertise. It's the marketing expertise, the systems expertise.