Decarbonization will require a completely transformed electricity grid. We analyze a long-run model that captures crucial aspects of the electricity industry such as time-varying demand for electricity, intermittency of renewables, optimal use of storage technologies, and entry and exit of generation and storage capacity. Long-run effects can differ in surprising ways from short-run intuition: A carbon tax can increase electricity consumption; cheaper storage may lead to a decrease in renewable capacity by decreasing prices when renewables are operating; and an increase in electricity demand (e.g., electrification) may lead to a decrease in emissions if it induces sufficient renewable entry. Using hourly data for the U.S. market, we calibrate the model to evaluate decarbonization policies. A carbon price of $150 or more essentially eliminates carbon emissions. Given a modest decarbonization goal, a renewable subsidy performs better than a nuclear subsidy, but this ranking is reversed for an ambitious decarbonization goal. Policies promoting transmission or storage are unlikely to yield significant benefits unless paired with subsidies for renewables. Electrifying 100% of car miles traveled (thereby eliminating gasoline vehicle carbon emissions) would increase electricity-sector carbon emissions by 23-27% if vehicles are charged at night, but could decrease electricity-sector carbon emissions if vehicles are charged during the day.