How does a virus bud from the plasma membrane of its host? Here we investigate several possible rate-limiting processes, including thermal fluctuations of the plasma membrane, hydrodynamic interactions, and diffusion of the glycoprotein spikes. We find that for bending moduli greater than 3 x 10(-13) ergs, membrane thermal fluctuations are insufficient to wrap the viral capsid, and the mechanical force driving the budding process must arise from some other process. If budding is limited by the rate at which glycoprotein spikes can diffuse to the budding site, we compute that the budding time is 10-20 min, in accord with the experimentally determined upper limit of 20 min. In light of this, we suggest some alternative mechanisms for budding and provide a rationale for the observation that budding frequently occurs in regions of high membrane curvature.