Dai and Karato (2014a) reported new electrical conductivity measurements on oriented single crystals of hydrous olivine showing a higher increase in conductivity and conductivity anisotropy at high temperature than previously reported. After recalling that the highest conductivities of the upper mantle (e.g. >0.1 S/m) cannot reasonably be explained by previous hydrous olivine conductivity data, they concluded from their new measurements that olivine hydration can actually account for the electrical anomalies of the upper mantle. However, we analyze the data of Dai and Karato (2014a) and demonstrate that they are not higher but lower than most previous measurements and models, including those of the same authors, even when extrapolated at upper mantle temperatures. This new study therefore does not extend but restricts the magnitude of olivine conductivity enhancement by water. We show that the high temperature conduction regime reported by Dai and Karato (2014a) cannot result from the diffusion of vacancy–hydrogen complexes as proposed by the authors, but is compatible with the diffusion of interstitial protons. We show that the high conductivity anisotropy observed in their olivine samples with high water concentration (∼1000 wt.ppm) must be drastically reduced at mantle-relevant water concentrations (View the MathML source<∼100 wt.ppm) since dry olivine is weakly anisotropic. We conclude from all of the existing models that olivine with mantle-relevant water concentrations is unlikely to produce significant conductivity anisotropy in deformed mantle and is unable to produce the highest conductivities of the upper mantle, which therefore likely result from partial melting.