We estimate the impact of ultra-fast broadband on schools' academic performance using a difference-indifference study of a new fibre broadband network. We show that fibre broadband increases primary schools' passing rates in standardised assessments by roughly one percentage point. Estimates are robust to alternative specifications, such as controlling for time-varying covariates. We find no evidence that gender, ethnic minorities or students enrolled in remote schools benefit disproportionately. However, we find some evidence of a larger benefit within schools that have a greater proportion of students from lower socioeconomic backgrounds.