This paper proposes the notion of territorial patterns of innovation as a new conceptualization to understand the relationship among knowledge, innovation, and economic growth at the regional level. The territorial patterns of innovation approach reject the simplistic view of an invention-innovation equivalence and advance alternative patterns, alternative ways in which knowledge and innovation can take place and mix in space. Each of them represents a different way of innovating, one not necessarily more efficient than the other. On the empirical ground, the paper demonstrates this statement for all NUTS2 regions of the 27 European Union states by showing that the efficiency in taking advantage of innovation does not only link to the strength of the local knowledge base; rather, territorial patterns of innovation characterized by relatively low knowledge intensity can be relatively more efficient in grasping and exploiting innovation returns for growing. Interesting policy implications can be drawn from the empirical analysis presented. If the results do not deny the importance of research and development (R&D) activities for regional growth, and therefore the right focus put forward by the Europe 2020 on a “smart growth” based on knowledge and innovation, they call for particular attention when the Europe 2020 goal is translated into a regional setting.