Purpose: To investigate the efficacy of intravenous GSK933776, a humanized monoclonal antibody directed against the N-terminal amino acids of amyloid β, for the treatment of geographic atrophy (GA) in age-related macular degeneration (AMD). Design: Prospective, randomized, placebo-controlled, double-masked, multicenter phase 2 clinical trial. Participants: Patients with GA secondary to AMD, a visual acuity score of at least 35 letters, and GA with a total area of 1.9 to 17 mm2 were enrolled. Methods: Participants were monitored monthly for 4 months during an observation period to determine the rate of GA enlargement in the study eye. After the observation period, randomization was performed into 1 of 4 treatment arms (GSK933776 at 3, 6, and 15 mg/kg/month and placebo). At each monthly visit over 18 months, participants underwent visual acuity testing under normal luminance and low-luminance conditions. Ocular imaging included color fundus photography, fundus autofluorescence, fluorescein angiography, and spectral-domain OCT. Main Outcome Measure: Enlargement in the area of GA measured from color fundus photographs with reference to fundus autofluorescence images. Results: A total of 191 participants were randomized into the study, with 139 (73%) fulfilling the efficacy population criteria. Over 18 months, GSK933776 did not reduce the rate of GA enlargement relative to placebo. Overall, there were no consistent meaningful differences relative to placebo in any of the visual function measures. There was a correlation between the low-luminance visual acuity (LLVA) deficit at baseline and the rate of GA enlargement. Genetic variations in complement factor I (CFI) gene did not correlate with GA progression. No ocular serious adverse events considered related to the GSK933776 treatment were identified, and a similar number of nonocular serious adverse events were reported across all treatment groups. Conclusions: Intravenous amyloid β inhibition with GSK933776 did not slow the rate of GA enlargement compared with placebo, and no clinically meaningful differences relative to placebo were observed in visual function testing over 18 months. The LLVA deficit was associated with faster GA enlargement; however, no correlation was shown between genetic variations in the CFI gene and the rate of GA enlargement.