FACULTY AND RESEARCH

We still poorly understand why Silicon Valley has originated so many breakthrough innovations and large companies. The durability of Silicon Valley's innovative competence over the last seventy years also needs more explanation. The failure of several policy-makers around the world to reproduce the Silicon Valley cluster reveals the misunderstanding of the innovative dynamic in Silicon Valley. This study uses complex network theory – CNT (Barabási, Newman & Watts, 2006; Jen, 2006; Thompson, 2004a) to analyse the complex innovative capability of Silicon Valley and to understand the heterogeneity of agents and the multiplexity of ties that support creation and development of high-tech start-ups. As proposed by Barabási (2002, p. 200), we view the economy as a complex network, whose nodes are companies and whose links represent the various economic and financial ties connecting them. Innovation and entrepreneurship are understood as resulting from the interactions of numerous economic agents. In a systemic perspective, by definition, the presence of a specific agent in a network induces specific interactions with other agents that could not take place if this agent were not there. Thus, the diversity of agents influences the dynamics of the system. The presence of venture capital (hereafter VC) firms in an innovative cluster opens potential specific interactions with other agents in the network (universities, large companies, laboratories) that determine a particular dynamic of innovation. In this perspective, what is distinctive about Silicon Valley is its complete and robust complex system of innovation supported by social networks of interdependent economic agents in which the VC firms have a specific function. Our perspective examines five different contributions of VC firms to Silicon Valley: financing, selection, collective learning, embedding and signalling. These five functions are different ways for the VC firms to interact with the other members of the complex network of innovation and to support the robustness of the system.