High clouds produced by tropical convection are expected to shrink in area as climate warms, and the radiative feedback associated with this change has long been the subject of controversy. In a recent assessment of climate sensitivity, the World Climate Research Programme (WCRP) estimated that the feedback is significantly negative, albeit with substantial uncertainty. Here we show that such a negative feedback is not supported by an ensemble of high-resolution atmospheric models. Rather, the models suggest that changes in cloud area and opacity act as a modest positive feedback. The positive opacity component arises from the disproportionate reduction in the area of thick, climate-cooling clouds relative to thin, climate-warming clouds. This suggests that thick cloud area is tightly coupled to the rate of convective overturning-which is expected to slow with warming-whereas thin cloud area is influenced by other, less-certain processes. The cloud response is examined from a novel perspective that treats high clouds as part of an optical continuum rather than entities with fixed opacity. The positive feedback differs significantly from previous estimates and leads to a 0.3 • C increase in climate sensitivity relative to a previous community assessment.