Should broadband subsidies make service affordable or extend networks to unserved areas? I compare the Affordable Connectivity Program (ACP) with the Broadband Equity, Access, and Deployment (BEAD) program using a new nationwide dataset linking prices, speeds, providers, and exposure across more than 70,000 census tracts. Continuous-treatment event studies show that ACP exposure reduced posted prices by 7–10 percent and shifted menus toward faster plans. BEAD effects appear after construction lags, when market structure shifts consistent with entry. I estimate a structural equilibrium model of endogenous prices, menus, and entry. Non-dominance inequalities bound fixed costs without solving a combinatorial entry game. At a 75 percent subsidy, BEAD generates 6,700–8,600 firm-market entries and about $7 in five-year welfare per public dollar, with gains largest in monopoly markets. ACP generates about $1 per dollar, concentrated among low-income households. Results imply sequencing: deploy where entry is unprofitable, then subsidize affordability where prices remain binding.