To examine whether suitable conditions occur for the water splitting reaction at their interfaces with liquid water, we determine the pH-dependent surface coverage for a series of semiconductors, including GaAs, GaP, GaN, CdS, ZnO, SnO2, and rutile and anatase TiO2. For this, we calculate acidity constants at surface sites through ab initio molecular dynamics simulations and a Grand Canonical formulation of adsorbates. The resulting pH values at the point of zero charge show excellent agreement with experiment and thereby support the validity of our approach. By combining information concerning the surface coverage with the alignment of the band edges with respect to the relevant redox levels, we scrutinize the potential of the considered semiconductors as photocatalysts and identify the corresponding optimal pH ranges for hydrogen and oxygen evolution. More specifically, our results indicate that GaN stands out among these semiconductors as the most promising candidate for the overall water splitting, with the potential of further improvement through alloying. With the surface coverage, our computational analysis brings a new descriptor that is currently beyond experimental reach.