A consensus has emerged about an incipient fourth industrial transition, which is being ushered in by pervasive technologies such as artificial intelligence (AI) and machine learning. This paper considers whether a new industrial transition needs a new innovation policy to guide the economic performance of cities and regions. In particular, the paper analyzes which innovation policies proved to be effective during the previous three industrial transitions and draws out insights that might suggest approaches likely to be effective in Industry 4.0. After introducing a framework identifying the forces or pillars underlying innovation and economic performance, policy strategies for industrial transition are considered, such as the diffusion of innovation and productivity, the external and internal factors influencing the efficacy of innovation policies, and the relative effectiveness of prioritizing specific or broad-based target.