AbstractThis paper proposes a model of the choice to commercialize research, of the amount and type of pre‐commercial research to perform, and of the timing of commercialization by an academic scientist, and analyzes the returns and costs of these choices. The behavior and performance of the academic scientist is compared with that of an industrial researcher. Unlike the industrial researcher, the academic scientist receives direct benefit from performing research, e.g. in the form of publication and peer recognition. However, the type of research that is more effective in reducing commercialization costs may not be the one generating the highest scientific benefit. It is shown that, while in some cases the academic scientist is more reluctant to commercialize research, in other cases she may commercialize faster than a solely profit‐seeking agent would—and perform less research. Academic and non‐academic scientists also select different projects, and this may explain the good performance of academic entrepreneurs found in several empirical studies. The model offers a unified framework to interpret the mixed evidence on the success of, and the arguments in favor and against, the involvement of universities into commercial activities. Managerial and public policy implications are also examined. Copyright © 2009 John Wiley & Sons, Ltd.