Based on our US-Japan comparative research on development of a blockbuster drug, namely, Statin, we investigate what kind of innovation system would facilitate the type of innovation shouldering high social risks. In this paper, we elucidate the hypothesis that commercialization of a newly developed dug can be explained by the pattern and function of researcher-networks among university, industry and public R&D institutes built in response to the emergence of new scientific discovery. Comparing the type of researcher-networks in US and Japan, the paper sheds light on roles of the networks for information sharing and consensus making, enabling us to conclude how the networks contribute to the success of high-risk innovations.