In an era of rapid climate change, and with it concern over insect declines, we used two theories to predict 20-year changes in 34 North American ant communities. The ecosystems, from deserts to hardwood forests, were first surveyed in the 1990s. When resurveyed in 2016-2017 they averaged 1°C warmer with 200 gC/m /y higher plant productivity. Ant colony abundance changed from -49% to +61%. Consistent with Thermal Performance Theory, colony abundance increased with temperature increases <1°C, then decreased as a site's mean monthly temperature change increased up to +2.4°C. Consistent with Species Energy Theory: 1) ant abundance tracked changes in a measure of energy availability (Net Aboveground Productivity, gC/m /y), and, 2) increases in colony abundance drove increases in local plot- and transect-level species richness but not that of Chao 2, an estimate of the size of the species pool. Even after accounting for these drivers, local species richness was still higher ca. 20 years after the original surveys, likely due to the increased activity of ant workers. These results suggest community changes are predictable using theory from geographical ecology, and that warming can first enhance but may ultimately decrease the abundance of this important insect taxon.