Abstract We perform direct numerical simulations to verify the hypothesis of invariance of the turbulence dissipation rate, proposed by Deane, Stokes, and Callaghan (2016, https://doi.org/10.1175/jpo‐d‐14‐0187.1) based on their laboratory breaking wave experiments. Our DNS data provide direct access to local viscous dissipation rates, which are difficult to measure in laboratory and field experiments. The breaking waves simulated have a wavelength of 25 cm and varying amplitudes. Despite a fivefold variation in the total energy dissipation, our results indicate the emergence of a space–time‐averaged dissipation rate between 0.49 and 0.70 W/kg. The additional energy loss accompanying an increase in wave amplitude is accommodated by an increase in the bubble plume volume, where energy dissipation is concentrated, rather than by an increase in energy dissipation rate via shear stress correlations. These findings clarify how breaking wave energetics constrain turbulence injection into the upper ocean and regulate air–sea exchange processes.