With its high degree of autonomy, dedication, and co-location, an autonomous team as an emerging tool for new product development (NPD) has more freedom and stronger capabilities to be innovative and entrepreneurial. However, such teams are not a panacea, and implementing them can be costly and disruptive to their parent organization. Hence, in this study we ask under which circumstances an autonomous team is the best choice for NPD.Drawing on both contingency and information-processing theories, we hypothesized that autonomous teams are more effective in developing novelty technology or radical innovation. We tested and confirmed the hypotheses using data from 555 NPD projects by comparing the relative effectiveness of autonomous teams with functional, lightweight, and heavyweight teams in terms of development cost, development speed, and overall product success. The results also suggest that heavyweight teams perform better than other teams in developing incremental innovation. The findings of this study may not only have some important implications for NPD practices but also shed some light on other important topics such as disruptive innovation, new venture, corporate entrepreneurship, and ambidextrous organization.