How should governments divide broadband subsidies between consumer discounts and infrastructure grants? I study the Affordable Connectivity Program (ACP) and the Broadband Equity, Access, and Deployment (BEAD) program, combining state-year event studies with a tract-level structural model of demand, pricing, and entry estimated on 170,419 provider-tier products in 66,454 census tracts built from FCC, ACS, and NTIA data. In high-eligibility states, ACP was followed by lower prices—roughly 7–10 percent at peak response—a shift toward faster plans, and lower concentration; BEAD effects are smaller and delayed, consistent with construction lags. Counterfactual simulations imply a BEAD benefit-cost ratio of roughly 26–27 discounted over the infrastructure's lifetime and an ACP ratio of one under targeted eligibility. ACP's returns concentrate where broadband exists but remains unaffordable; BEAD's where infrastructure is missing. The optimal allocation depends on the binding constraint: affordability in served low-income markets, deployment in unserved ones.