Progress to implement and assess an implicit solution capability in the spectral element dynamical core of the Community Atmosphere Model is presented. Within the testbed shallow water version of the model, efficiency gains and scalability are demonstrated through effective data transfer and a SIMPLE-based preconditioner. In the full primitive equation model, using implicit methods reduces the levels of subcycling in the model. Because numerical time error is small relative to other model sensitivities, implicit methods provide an efficiency gain through much larger time steps. Several examples of success and limitations are presented.