One tenth of anthropogenic greenhouse gas emissions are caused by heating and cooling buildings. Efficient electric heat pumps could significantly reduce these emissions, but face barriers to adoption related to costs, equipment selection and installation, and other factors. The goal of this thesis is to reduce emissions by lowering barriers to heat pump adoption. To this end, we investigate heat purchase agreements (HPAs), a new model of heat pump ownership, and develop supporting methods. In an HPA, users host heat pumps owned by an aggregator. The aggregator buys the heat pumps' electricity and sells their heat or cooling output to the users. We show that HPAs can lower barriers to adoption and benefit both the aggregator and the users. We also develop a method for fairly pricing heat and cooling. An HPA aggregator is responsible for selecting an appropriate heat pump for each user under uncertainty. We develop a data-driven selection method that provides probabilistic feasibility and optimality guarantees, and illustrate the method through simulations. An HPA aggregator operates a fleet of heat pumps. If the aggregator invests in sensing, communication and control capabilities, then they can provide services to the electricity grid by perturbing the heat pumps' power use. We develop methods for co-optimizing day-ahead capacity offers for the two highest-priced services, regulation and spinning reserve. In simulations, each heat pump offers 285--325 W of combined annual-average capacity and earns $25--75 of annual revenue. Providing these services could help grid operators integrate more renewable power, and thereby reduce emissions from electricity generation.