I examine how two different realizations of bed morphology affect Thwaites Glacier response to ocean warming through the initiation of marine ice sheet instability and associated grounding line retreat. A state of the art numerical ice sheet model is used for this purpose. The bed configurations used are the 1-km resolution interpolated BEDMAP2 bed and a higher-resolution conditional simulation produced by John Goff at the University of Texas using the same underlying data. The model is forced using a slow ramp approach, where melt of ice on the floating side of the grounding line is increased over time, which gently nudges the glacier toward instability. Once an instability is initiated, the anomalous forcing is turned off, and further grounding line retreat is tracked. Two model experiments are conducted. The first experiment examines the effect of different anomalous forcing magnitudes over the same bed. The second experiment compares the generation and progress of instabilities over different beds. Two fundamental conclusions emerge from these experiments. First, different bed geometries require different ocean forcings to generate a genuine instability, where ice dynamics lead to a positive feedback and grounding line retreat becomes unstable. Second, slightly different forcings produce different retreat rates, even after the anomalous forcing is shut off, because different forcing magnitudes produce different driving stresses at the time the instability is initiated. While variability in the retreat rate over time depends on bed topography, the rate itself is set by the magnitude of the forcing. This signals the importance of correct knowledge of both bed shape and ocean circulation under floating portions of Antarctic ice sheets. The experiments also imply that different ocean warming rates delivered by different global warming scenarios directly affects the rate of Antarctic contribution to sea level rise.