Observations in tectonically active areas increasingly reveal sets of high-angle conjugate faults (‘cross-faults’) that apparently contradict theories of faulting based on experimental data. Possible explanations include a low in situ coefficient of friction, dominant control of ductile shear zones in the lower crust, and tectonic rotation. Discrimination between these mechanisms has been hindered by uncertainties in the state of stress, deformation history and fault geometry at seismogenic depths. Here we use a combination of seismic, geodetic and geologic data to demonstrate that ubiquitous cross-faults in the Ridgecrest (California) area, including those ruptured in the sequence of strong earthquakes in 2019, result from rotation from an ini- tially optimal orientation consistent with experimental data. The inferred rotation pattern can be explained by the geodetically measured velocity field. Our model suggests that the observed asymmetric rotation of faults in the Eastern California Shear Zone can result from simple shear. The same mechanism can be responsible for high-angle conjugate faults observed in other tectonic settings.