2
A Lesson About Precompetitive Collaboration
Key Points Raised by the Speaker
|
THE SEMATECH EXPERIENCE
William Spencer, the chairman emeritus of SEMATECH, described how the semiconductor manufacturing industry responded when it faced a challenge that required collaborative efforts among competing companies.
In 2010 the U.S. semiconductor industry should make about $300 billion in sales, said Spencer, while supporting an electronics industry that does about a trillion and a half dollars in business per year. “If you count the number of transistors you have on yourself right now—with your cell phone, your pager, your watch, the key to your car, some of your credit
cards, and if you have a laptop with a couple of memory sticks in it—you’ve probably got 10 [billion] to 100 billion of those little buggers on your body or in your purse right now. That’s close to the number of cells that you have in your body, not quite the number, but the transistors are multiplying faster,” said Spencer.
The transistor was invented by two scientists at Bell Laboratories in 1947 (Brinkman, 1997), and about a decade later two separate companies, working independently, discovered how to combine many transistors onto a single chip of silicon. This started the semiconductor industry whose growth, according to Spencer, took off in the 1960s when the U.S. government chose to use integrated circuits in military systems, including the Minuteman missile. At that time, the semiconductor industry was small—perhaps just a billion dollars—and the Department of Defense accounted for about half of those sales.
By the mid-1970s, the industry had grown to approximately $5 billion in sales, and integrated circuits were being used much more widely. At that time, the United States owned about two-thirds of the world market in semiconductor chips and almost 100 percent of the market in the complex equipment used to manufacture those chips. However, other countries, particularly Japan, saw an advantage in semiconductor manufacturing and in 1975 started a cooperative effort between industry and government to boost the industry (Sakakibara, 1993).
According to Spencer, by 1985, the United States had lost market leadership in semiconductor manufacturing. At that point, it cost about a billion dollars to build a new factory to make semiconductor chips, and the equipment for a factory might cost several hundred million dollars. If a company made a mistake in building a plant, failure represented a huge cost, but if a company chose not to invest in plants it lost market share and future revenue.
The Department of Defense did not think that buying semiconductor chips from outside the United States was a good idea due to national security concerns about reliable and secure sources of chips, according to Spencer. This concern contributed to the formation of SEMATECH—an acronym for semiconductor manufacturing technology—as a cooperative effort between the federal government and the U.S. semiconductor industry. 1 SEMATECH initially had a budget of about $200 million, representing about 75 percent of the U.S. semiconductor industry, with a little less than 50 percent of that amount coming from the federal government and the rest from private industry (CBO, 1990).
At first, former competitors working together at SEMATECH were wary. Spencer recalled an engineer telling him that her company had told her to “listen, don’t talk, and lock your file cabinet.” However, the work at SEMATECH quickly revealed the advantages of cooperation. The problems being tackled at SEMATECH were being worked on independently by different companies. “By getting together, those problems could be solved jointly, saving a great deal of time and a great deal of money,” observed Spencer.
By the mid-1990s, the United States had regained its market leadership in the semiconductor industry. There were many reasons for that, Spencer acknowledged, including changes in tariffs, exchange rates, trade barriers, and technologies, but SEMATECH clearly played a role.
LESSONS FROM SEMATECH
Spencer highlighted several key points from his experience at SEMATECH that were necessary for its success. First, it is important to have stable funding for such a venture. Funding for SEMATECH was provided through regular federal appropriations and industry. Federal support for SEMATECH was in the President’s budget that was submitted each year. Inclusion in the President’s budget is critical, he said, to avoid budget battles within each agency.
Second, according to Spencer, industry-led collaborations need ties to universities to draw on the creative ferment of academia. A close association is advantageous for the initiation of cooperative efforts with faculty who are already experts in disparate fields. Additionally, it allows for identification of outstanding graduates who could work directly for SEMATECH or one of the member companies. Spencer noted that while a consortium might be located close to a specific university, it is also important to establish relationships with every other university where there is an overlapping common interest.
Most important, said Spencer, such ventures need to attract the best people to be successful. To do that, those people need to see a way to enhance their careers. Companies needed to be convinced that they should send their best people to SEMATECH and that once those individuals returned to the company, they should receive some kind of promotion, bonus, or recognition. If companies did not do that, their best people tended to leave for other companies, since the best people in the industry were widely recognized and sought by competing companies. “They could send them to SEMATECH and they’d return a lot smarter and a lot more knowledgeable than they were when they left,” said Spencer.