Riverine heatwave events—periods of abnormally high riverine water temperatures—can substantially impair aquatic ecosystems, water quality, and food and energy production. Large-scale analysis of riverine heatwaves however has been hindered by fragmented and discontinuous water temperature data. Here we used a deep learning model and reconstructed consistent daily water temperatures in 1471 U.S. River sites (1980–2022). Results show that riverine heatwave events occur less frequently and intensively but last nearly twice as long as air heatwaves. Alarmingly, riverine heatwaves have accelerated at faster rates than air heatwaves. Results here underscore the need for coordinated monitoring and data consolidation efforts for riverine heatwaves, and their incorporation into global climate risk assessment and adaptation policies.