Cool and green roofs are widely adopted measures for curtailing summertime urban heat islands. Existing numerical studies to assess their effectiveness and cooling benefits usually assume an unrealistic 100% coverage across the entire metropolis. This study investigates the scale dependence of the absolute cooling benefits and efficiency (cooling per adapted roof area) of cool and green roofs in a typical summer when they are installed over 25% of building rooftops at local, city, or regional scales. Six major U.S. cities with active climate action plans in different geoclimatic zones are compared through high-resolution simulations using the Weather Research and Forecasting model. The results reveal that reductions in 2-m air temperature over the urban core increase non-linearly with the intervention area, and the benefits of both roof types scale similarly. This scale-dependence of urban core cooling is not universal, but is rather controlled by the shape of metropolitan areas and wind pattern. The siting of mitigation measures hence plays an important role especially under windy conditions, and some urban cores are not able to achieve a noticeable and consistent cooling by retrofitting their own rooftops. Regional-scale deployments of mitigation strategies, on the other hand, yield a more substantial temperature reduction but with a lower efficiency. The scale-dependence of regional cooling efficiency showed remarkable similarity across studied cities, yielding a potentially generalizable power law. The successful resiliency plans for cities should account for the scale dependence and geoclimatic determinants of the achievable cooling, and identify the target neighborhoods of most interest.