Introduction: Animal models of Alzheimer's disease show that exercise may modify β-amyloid (Aβ) deposition. We examined the effect of a 16-week exercise intervention on cortical Aβ in patients with mild-to-moderate Alzheimer's disease. Methods: Thirty-six patients with Alzheimer's disease were randomized to either one hour of aerobic exercise three times weekly for 16 weeks or usual care. Pre and post intervention, 11Carbon-Pittsburgh compound B positron emission tomography was carried out to assess cortical Aβ, and quantified using standardized uptake value rations (SUVRs). Results: The intervention showed no effect on follow-up SUVRs in a covariance analysis with group allocation, baseline intervention SUVR, age, sex, and baseline Mini–Mental State Examination as predictors. Change in SUVRs did not correlate with changes in measures of physical or aerobic fitness. Discussion: The present findings do not support an effect of exercise on Aβ. However, the relatively short intervention period may account for a lack of efficacy. Further studies should test earlier and longer interventions.