This paper investigates the effect of accessing ultra fast broadband (UFB) on firm performance. To address endogeneity issues, we focus on a French policy deploying UFB in areas for which the private sector (telecommunication companies) expressed no interest in deploying the technology. Given the top down nature of the policy and the important uncertainties regarding the delays of the deployment, we argue that the timing of access to UFB for the companies is plausibly exogenous. The areas affected by this policy represent about 30,000 cities and 43% of the French population. Thanks to granular firm-level data and precise information on internet access, we leverage the timing of access to draw causal estimates from a staggered difference in difference design at the city level. We show that UFB access increases value added of the firms in our sample right from the year of access and this positive effect grows over time. This increase in value added comes from companies existing before the arrival of the UFB and not from companies arriving after (i.e. not from ex post inward mobility) nor from exists of low productivity firms. Regarding employment, we evidence an increase in employment but only after 4 years. We find no effect of UFB on labour productivity.