Abstract Adiabatic compressed air energy storage (ACAES) is an energy storage technology that has the potential to play an important role in the transition to a predominantly renewables‐driven net‐zero energy system. However, it has not yet achieved the performance necessary to be widely deployed. This paper undertakes an exergy analysis of isobaric and isochoric ACAES systems, tracking lost work through the components and exploring the influences of different design choices. Three different configurations are modelled: (1) 3 compression and 3 expansion stages; (2) 4 compression and 2 expansion stages; and (3) 2 compression and 4 expansion stages. These results illustrate that isobaric systems are likely to have higher round‐trip efficiency and significantly higher energy density, at the cost of achieving isobaric storage. Exergy analysis reveals that most of the losses arise in the compressors, compressor aftercoolers and expanders. Losses in aftercoolers are exaggerated when compressors operate with high‐pressure ratios, emphasizing that the choice of TES is a key system variable. With pressurised water as the coolant and TES fluid, it seems likely that the best system will have more compression than expansion stages. Increasing the number of compression stages decreases the off‐design penalty when the system is isochoric.