Background Shoulder injuries constitute a considerable risk to professional rugby union players; however, there is a shortage of detailed epidemiologic information about injuries in this population. Purpose To describe the incidence, severity, and risk factors associated with shoulder injuries in professional rugby union. Study Design Descriptive epidemiology study. Method Medical personnel prospectively reported time-loss injuries in professional rugby union in England, and the shoulder injuries were evaluated. Results The incidence of shoulder injuries was significantly lower during training (0.10/1000 player—training hours) compared with matches (8.9/1000 player—match hours). The most common match injury was acromioclavicular joint injury (32%); the most severe was shoulder dislocation and instability (mean severity, 81 days absent), which also caused the greatest proportion of absence (42%) and had the highest rate of recurrence (62%). The majority of match shoulder injuries were sustained in the tackle (65%), and outside backs were the most likely to sustain an injury from tackling (2.4/1000 player-tackles). Injuries sustained during training were significantly more severe (61 days) than were those sustained during match play (27 days), and defensive training sessions carried the highest risk of injury (0.45/1000 player-hours; mean severity, 67 days). A mean of 241 player-days per club per season were lost to shoulder injuries. Conclusion Results suggest the potential to reduce this injury burden by modifying training activities and implementing “prehabilitation” strategies in an effort to minimize the risk of shoulder dislocation/instability.