BACKGROUND Glenoid component loosening is common in total shoulder arthroplasty (TSA), often resulting from the mechanical interaction of glenohumeral components. This cadaveric study was performed to evaluate and to compare commercially available onlay and inlay glenoid prosthetic designs with respect to loading characteristics and loosening. METHODS Sixteen prescreened cadaveric shoulders (8 matched pairs) underwent either onlay or inlay TSA. We created a custom glenohumeral loading model and used cycles of 5 mm anterior-posterior humeral translation to simulate a rocking-horse loosening mechanism for all testing. Articular TekScan measurements were performed with 9.1 kg (88.9 N) of glenohumeral compression before and after TSA. Fatigue testing was performed with 34.0 kg (333.6 N) of glenohumeral compression using high-definition video to document gross glenoid loosening. Testing ended with gross loosening or a maximum of 4000 cycles. Mean contact area, pressure, and joint reaction force were used to compare the 2 glenoid designs. RESULTS In both implant types, contact area decreased and pressure increased after TSA (P < .0001). Force increased at the onlay component edge only (P = .0012) compared with native glenoid testing. Force was greater in the onlay vs. the inlay implants (P < .0001). During fatigue testing, all onlay glenoid components exhibited gross loosening at a mean of 1126 cycles (range, 749-1838), whereas none of the inlay glenoid components exhibited gross loosening (P < .0001). CONCLUSION The inlay glenoid implant exhibited biomechanical characteristics favoring stability and decreased loosening compared with the onlay glenoid implant in this cadaveric model.