Globally, heat stress (HS) is nearly certain to increase rapidly over the coming decades, characterized by increased frequency, severity, and spatiotemporal extent of extreme temperature and humidity. While these characteristics have been investigated independently, a holistic analysis integrating them is potentially more informative. Using observations, climate projections from the CMIP5 model ensemble, and historical and future population estimates, we apply the IPCC risk framework to examine present and projected future potential impact (PI) of summer heat stress for the contiguous United States (CONUS) as a function of non-stationary HS characteristics and population exposure. We find that the PI of short-to-medium duration (1-7 days) HS events is likely to increase more than three-fold across densely populated regions of the U.S. including the Northeast, Southeast Piedmont, Midwest, and parts of the Desert Southwest by late this century (2060-2099) under the highest emissions scenario. The contribution from climate change alone more than doubles the impact in the coastal Pacific Northwest, central California, and the Great Lakes region, implying a substantial increase in HS risk without aggressive mitigation efforts.