AbstractAcross the globe, temperatures are predicted to increase with consequences for many taxonomic groups. Arthropods are particularly at risk as temperature imposes physiological constraints on growth, survival, and reproduction. Given that arthropods may be disproportionately affected in a warmer climate—the question becomes which taxa are vulnerable and can we predict the supposed winners and losers of climate change? To address this question, we resurveyed 33 ant communities, quantifying 20‐yr differences in the incidence of 28 genera. Each North American ant community was surveyed with 30 1‐m2 plots, and the incidence of each genus across the 30 plots was used to estimate change. From the original surveys in 1994–1997 to the resurveys in 2016–2017, temperature increased on average 1°C (range, −0.4°C to 2.5°C) and ~64% of ant genera increased in more than half of the sampled communities. To test Thermal Performance Theory's prediction that genera with higher average thermal limits will tend to accumulate at the expense of those with lower limits, we quantified critical thermal maxima (CTmax: the high temperatures at which they lose muscle control) and minima (CTmin: the low temperatures at which ants first become inactive) for common genera at each site. Consistent with prediction, we found a positive decelerating relationship between CTmax and the proportion of sites in which a genus had increased. CTmin, by contrast, was not a useful predictor of change. There was a strong positive correlation (r = 0.85) between the proportion of sites where a genus was found with higher incidence after 20 yr and the average difference in number of plots occupied per site, suggesting genera with high CTmax values tended to occupy more plots at more sites after 20 yr. Thermal functional traits like CTmax have thus proved useful in predicting patterns of long‐term community change in a dominant, diverse insect taxon.