In this paper, we optimize the energy efficiency of device-to-multi-device (D2MD) communications. While point-to-point has been extensively studied to improve the spectral efficiency in cellular networks, the introduction of multicast communications brings, in addition, the possibility of reusing the spectrum resources also inside the groups. We formulate the energy efficiency problem in this context, and show that a decomposition approach which first uses matching theory to allocate the channels to each group, and then optimizes to find the unique transmission power vectors in each group gives good performance results for the optimization both of the system’s energy efficiency and of the max-min individual energy efficiency. This is achieved for a desired minimum transmission rate. Numerical results are presented to confirm the usefulness of the proposed framework for optimizing the planning and deployment of D2MD networks with energy efficiency as a primary performance goal.