The Pacific Earthquake Engineering Research Center Next Generation Attenuation‐West2 database is used to derive a new conditional ground‐motion model (CGMM) and a set of scenario‐based models for estimating cumulative absolute velocity (CAV) for earthquakes in shallow crustal tectonic settings. Random‐effects regressions were performed to develop the conditional model, with random effects across different earthquakes. The estimate of CAV is conditioned on the estimated peak ground acceleration (PGA), the time averaged shear‐wave velocity in the top 30 m (), the earthquake magnitude (), and the rupture distance (). By combining the conditional CAV model with ground‐motion models (GMM) in shallow crustal earthquake zones for PGA, new scenario‐based models are developed for estimating the median CAV and its standard deviation, directly from an earthquake scenario and site conditions. A scenario‐based CAV model captures inherently the complex ground‐motion scaling effects included in the GMMs for spectral accelerations on which it is based on, such as, sediment‐depth effects, soil nonlinearity effects, and regionalization effects. This approach also ensures consistency between the estimated CAV values and a design spectral acceleration response spectrum. The conditional and scenario‐based models to estimate CAV are presented, and trends of the developed scenario‐based models and previous traditional models for CAV are compared. Interestingly, we found a remarkable consistency between scenario‐based and traditional nonconditional CAV models, when the underlain spectral GMM used in the implementation of the scenario‐based model is properly constrained. Finally, we provide examples for the use of the conditional and scenario‐based models in performance‐based earthquake engineering.