The shattered pellet injection (SPI) system will be used at ITER as the primary disruption mitigation system (DMS) [1]. Experiments in present devices play a key role in understanding the plasma response to SPI. The SPI system in ASDEX Upgrade (AUG) [2] has been in operation since 2022 and has carried out various discharges with varying SPI and plasma parameters. In support of the experimental activities carried out at various tokamaks, additional modelling activities are also required for the physics-based extrapolation from the experimental results in present devices to ITER. For accurate modelling of the material assimilation, it is important to match the material penetration in the simulations with the experiments, especially considering the importance of two phenomena: the plasmoid drift and the rocket effect. The penetration can be inferred in AUG SPI experiments using a top-down view, visible range, fast camera. We discuss some of the key observations made so far in quantifying the material penetration for different pellet and plasma parameters. Statistical analysis of the 2022 and 2025 SPI experiments indicates that the speed of the ablation front, as observed from the top-down fast camera is generally less than 80\% of the pellet speed parallel to the injection vector, with minimal dependence on any relevant pellet parameters. Analysis of fast camera recordings for plasmoid drift suppression scans at AUG in 2025 also indicate no significant changes in the material penetration for slow or intermediate speed pellets while material assimilation measurements (from A. Moreau in this meeting) indicate higher material assimilation for pellets with neon doping compared to pure deuterium pellets. We also discuss the inclusion of the rocket force in our AUG SPI modelling efforts. [1] M. Lehnen, *et al.* TECH/1–1, 2023 - IAEA (2023) [2] M. Dibon, *et al.* Rev. Sci. Instrum. 94, 043504 (2023)