Ultrasonic bonding of low-temperature PEM membrane electrode assembly (MEA) components together has been shown previously to cut both cycle time and energy input of that manufacturing step by over an order of magnitude as compared to the industry standard of thermal pressing. This paper compares performance between ultrasonically and thermally bonded low-temperature MEAs and characterizes the performance losses from the new bonding process. A randomized, full factorial experiment was designed and conducted to examine performance of MEAs with 10 cm2 active area while varying three factors: bonding method (ultrasonically and thermally pressed using previously optimized bonding parameters), membrane condition (dry and conditioned Nafion® 115), and electrode catalyst loading (0.16 and 0.33 mg Pt/cm2). Ultrasonic MEAs performed as well as their thermal MEAs across all tested current densities with pure oxygen supplied to the cathode. However, thermal MEAs outperformed ultrasonic MEAs at current densities above 0.4 A/cm2 with air supplied to the cathode. Impedance spectroscopy, cyclic voltammetry, and flow sensitivity analyses were used to characterize the performance losses of the ultrasonic MEAs. The data suggest the presence of oxygen diffusion losses above 0.4 A/cm2 when air was supplied to the cathode. Ultrasonic MEAs were three times more sensitive to changes in air flow rate on the cathode than the thermally MEAs. Increasing the platinum catalyst loading from 0.16 to 0.33 mg Pt/cm2 resulted in a performance enhancement of approximately 20 mV and 65% greater electrochemical surface area. Understanding the effect of ultrasonic bonding on various performance losses will help optimize the MEA bonding process. Analysis of specific losses present for ultrasonic MEAs may also provide insight into the design of MEA components for ultrasonic bonding.