This paper combines data from a government programme providing broadband access to primary schools in Ireland with survey microdata on schools', teachers' and pupils use of the internet to examine the links between public subsidies, classroom use of the internet and educational performance. Provision of broadband service under a government scheme was associated with more than a doubling of teachers' use of the internet in class after about a two year lag. Better computing facilities in schools were also associated with higher internet use, but advertised download speed was not statistically significant. A second set of models show that use of the internet in class was associated with higher average mathematics scores on standardised tests, but that any association with reading scores was marginal. A range of confounding factors is also explored, with results broadly in line with previous literature.