The direct capture of CO2 from air (DAC) has been shown a growing interest for the mitigation of greenhouse gases but remains controversial among the engineering community. The high dilution level of CO2 in air (0.04%) indeed increases the energy requirement and cost of the process compared to carbon capture from flue gases (with CO2 concentrations around 15% for coal power plants). Until now, solid sorbents (functionalized silica, ion exchange resins, metal–organic frameworks, etc.) have been proposed to achieve DAC, with a few large-scale demonstration units. Gas-liquid absorption in alkaline solutions is also explored. Besides adsorption and absorption, membrane processes are another key gas separation technology but have not been investigated for DAC yet. The objective of this study is to explore the separation performances of a membrane unit for CO2 capture from air through a generic engineering approach. The role of membrane material performances and the impact of the operating conditions of the process on energy requirement and module production capacity are investigated. Membranes are shown to require a high selectivity in order to achieve purity in no more than two stages. The specific energy requirement is globally higher than that of the adsorption and absorption processes, together with higher productivity levels. Guidelines on the possibilities and limitations of membranes for DAC are finally proposed.