Photocatalysis using semiconductor powders in suspension performs reduction and oxidation reactions at nanometer-scale distances. Such short distances between the reduction (cathode) and the oxidation (anode) sites enable photocatalytic water splitting to generate H2 and O2 from pure water without a supporting electrolyte, which is otherwise impossible in conventional electrode systems due to the high solution resistance. A CrOx/Pt/SrTiO3 model photocatalyst achieves high efficiency under UV irradiation in ultra-pure water splitting at rates (>1 μmol-H2 per cm2 per h) corresponding to electrocatalysis on the order of mA cm−2. The introduction of an unbuffered supporting electrolyte did not improve the photocatalytic rates, consistent with the negligible ohmic losses (100 mV when rate >1 μmol-H2 per cm2 per h) even when the distance between redox sites was below 10 nm. Substantial improvements in photocatalytic rates were observed when buffer ions were introduced into near-neutral pH media by not only circumventing pH gradients but introducing kinetically facile H+ reduction to H2 instead of the kinetically sluggish direct reduction of H2O to H2. Herein, the quantitative descriptions of the electric potential, concentration gradients, and catalytic performance in nanoscale water electrolysis are presented with emphasis on (1) the advantages of performing redox reactions at the nanoscale, (2) the use of electrolyte engineering at near-neutral pH as a universal and effective strategy, and (3) the effectiveness of transferring knowledge from electrocatalysis to photocatalysis, where the potential is quantitatively defined regarding the former and poorly quantified regarding the latter.