The oxygen evolution reaction (OER) is the anodic half-reaction of water splitting, in which water is oxidised to molecular oxygen with the release of four protons and four electrons. It is the bottleneck of electrolytic hydrogen production: the reaction is a multi-step proton-coupled electron transfer whose intermediates bind to the surface through correlated energies, so a catalyst cannot be optimised for one step alone, and the overpotential required on even the most active materials is substantial. Because acidic operation demands iridium or ruthenium oxides, much of the work concerns earth-abundant catalysts that operate in alkaline medium, and the progress on bifunctional materials that also evolve hydrogen has been reviewed ^[1]. Cobalt-based catalysts are the most studied of these; dispersing the metal in a conductive carbon matrix raises the utilisation and gives a durable electrode ^[2], and nanoengineering the cobalt phase to expose more of the active site improves the intrinsic activity further ^[3]. Transition-metal oxides that are not themselves noble have also been examined, vanadium dioxide among them ^[4]. Compositionally complex alloys are explored as a way of tuning the adsorption energies continuously ^[5]. The surface of such catalysts is usually reconstructed under anodic bias, so the phase that is prepared is rarely the phase that operates.