Industries with strong complementarities and network externalities may exhibit multiple equilibria. I examine whether direct government provision of network infrastructure can remedy this inefficiency. Using car registration and charging transaction data from 2017–2023, I estimate a structural model of vehicle demand and charging behavior, where consumers choose an EV and derive utility from undertaking OD trips given the charging network topology. I decompose the Big Push into two components: a coordination effect and an externality effect. I find that the coordination effect — shifting the economy from the low Nash-Equilibrium to the high one — increases EV adoption by 140 %, while the externality effect raises adoption by a compounding 50 %. The latter arises from placing chargers at critical connector nodes that would otherwise be privately unprofitable. As the network grows from 300 to 1,000 stations, government provision crowds in private investment in the core while crowding it out at the periphery. Overall, the policy is highly effective, generating $11 billion in consumer value and $2,7 billion in climate benefits from a $185 million investment. This paper provides the first empirical evidence that direct public provision of network infrastructure can resolve coordination failures.