An analytic and numerical study of the feasibility of depressed collectors for gyrotrons for accelerator applications is discussed, and a specific design for a 10-GHz 30-MW gyroklystron is realized. The conclusion of the study is that depressed collectors are feasible for gyrotrons of interest for accelerator applications (i.e. P<or=100 MW, 10 GHz<f<35 GHz) and that their use would reduce capital costs for the RF system for a typical large accelerator by roughly 40%. In addition, scaling studies indicate that depressed collectors are feasible for fusion-related gyrotron oscillators, such as the 1-MW, 280-GHz devices being developed for use on the Compact Ignition Tokamak (CIT).<<ETX>>