Les secrets de la Silicon Valley ou les entreprises américaines de microélectronique face á l'incertitude technique

Semiconductor fi rms developed several organizational strategies to manage technological uncertainty since the 1950s. Over this period, they faced three types of technological uncertainty : cognitive uncertainty (how do devices and processes really work?), material uncertainty (how do electronic materials behave ?), and prospective uncertainty (where is microelectronics technology going ?). In the 1950s and early 1960s, they looked for the right balance between fundamental research and trial-and-error engineering to reduce these forms of technological uncertainty. As a new wave of innovation in microelectronics occurred in the second half of the 1960s, semiconductor fi rms clustered in industrial districts such as Silicon Valley to get ready access to the latest developments in microelectronics technology. Perhaps the most innovative intellectual and organizational tool that semiconductor engineering managers developed to manage technological uncertainty was in the area of technology forecasting : Moore's Law and the technology roadmaps for semiconductors. These tools, developed from the 1970s to the early 1990s, were largely directed to issues of prospective uncertainty. Moore's Law offered a useful method to guide technology investments. The pursuit of Moore's Law was later institutionalized by technology roadmapping exercises, particularly the National Technology Roadmap for Semiconductors. The roadmaps helped semiconductor manufacturers and their suppliers manage uncertainty and coordinate their research and development efforts. The organizational strategies initiated by the semiconductor industry to reduce technological uncertainty have been employed by other high tech sectors such as biotechnology, nanotechnology and the photovoltaic industry since the 1990s.