This paper empirically investigates the impact of broadband diffusion on crime. I employ state-level data on high-speed Internet penetration and criminal activity in the United States from 2000 to 2012. To tackle the endogeneity of broadband diffusion, I rely on a set of technologically motivated instrumental variables. The outcomes show that the impact of high-speed Internet diffusion on crime is negative and it is not significant for more severe offenses. Concerning the underlying mechanism, I find support for the hypothesis that the broadband-driven crime reduction occurs through a voluntary incapacitation effect, as people change their leisure time allocation by spending more time at home and off the streets. This in turn reduces the availability of criminal opportunities – and crime levels. Finally, my results suggest that the voluntary incapacitation effect concerns both sides of criminal interactions.