Predicting the final embedment depth of a dynamically installed anchor is a key prerequisite for reliable calculation of anchor capacity. This paper investigates the embedment characteristics of dynamically installed anchors in normally consolidated and overconsolidated clay through a series of centrifuge tests involving a model anchor instrumented with a microelectric mechanical system (MEMS) accelerometer, enabling the full motion response of the anchor to be established. The data are used to assess the performance of an anchor embedment model based on strain-rate-dependent shearing resistance and fluid mechanics drag resistance. Predictions of a database of over 100 anchor installations — formed from this study and the literature — result in calculated anchor embedment depths that are within ±15% of the measurements. An interesting aspect, consistent across the entire database, relates to the strain rate dependence on frictional resistance relative to bearing resistance. The predictions reveal that strain rate dependency may indeed be higher for frictional resistance, although only if a soil strength lower than the fully remoulded strength is considered as the reference strength, which suggests that water may be entrained along a boundary layer at the anchor–soil interface during installation.