Exercise is an emerging cause of atrial fibrillation (AF) in young individuals without coexisting cardiovascular risk factors. The causes of exercise-induced atrial fibrillation remain largely unknown, and conclusions are jeopardised by apparently conflicting data. Some components of the athlete's heart are known to be arrhythmogenic in other settings. Bradycardia, atrial dilatation and, possibly, atrial premature beats are therefore biologically plausible contributors to exercise-induced AF. Challenging findings in an animal model suggest that exercise might also prompt the development of atrial fibrosis, possibly due to cumulative minor structural damage after each exercise bout. However, there is very limited, indirect data supporting this hypothesis in athletes. Age, sex, the presence of comorbidities and cardiovascular risk factors, and genetic individual variability might serve to flag those athletes who are at the higher risk of exercise-induced AF. In this review, we will critically address current knowledge on the mechanisms of exercise-induced AF.