Future climate warming is virtually certain to bring about an increase in the frequency of heat extremes. Highway design and pavement selection are based on a temperature regime that reflects the local climate zone. Increasing heat extremes could, therefore, shift some areas into a different pavement grade (PG), and more heat-resistant materials are associated with increased infrastructure costs. This study combines observations, global climate model output, and a statistical model to investigate changes in 20-year return values of extreme maximum pavement temperature (TPmax). From a multimodel range of simulated TPmax, future changes in PG are computed for 17 major Canadian cities. Relative to a 1981-2000 baseline, summertime Canada-wide warming of 1-3°C is projected for the period 2041-70. As a result, climate change is likely to bring about profound changes to the spatial distribution of PG, with the severity of the changes directly linked to the severity of the projected warming. Even under weak simulated warming, an increase in PG is projected for Canada’s largest urban area, the Greater Toronto Area, and under moderate (strong) warming 7 out of 17 (9 out of 17) major cities exhibit an increase. The influence of model spatial resolution is evaluated by comparing the results from global climate models with output from a set of regional climate models focused on North America. With the exception of mountainous terrain in western Canada, spatial resolution is not a major determining factor for projections of future PG changes.