In recent years, Diecast has emerged as a transformative technology in automotive manufacturing, enabling the integration of numerous components—potentially reducing 171 individual parts into a single large-scale casting. While this approach significantly minimizes assembly complexity and production costs, it introduces critical challenges regarding structural reliability. The presence of porosity leads to substantial uncertainty in stress-strain curves (SS curves). This variability hinders the precise performance guarantees required for safety-critical automotive applications. To address this, we propose a computational framework to probabilistically characterize the mechanical response of Diecast structures by constructing a large-scale synthetic dataset of tensile analyses. Our approach utilizes Eulerian structural simulation, which is highly suited for high-fidelity modeling of structures containing porosity due to its superior parallelization efficiency and the elimination of complex mesh generation for internal geometries. Furthermore, to ensure the physical relevance of the simulated defects, we employ a 3D generative AI model based on DeepSDF. This model is trained on experimental data from real porosity to synthesize realistic, stochastic porosity patterns. By performing iterative Eulerian structural simulations on structures embedded with these AI-generated porosity, we aim to quantify the probabilistic distribution of SS curves. In this study, we present the foundational validation of this framework. We demonstrate that our Eulerian structural simulation accurately reproduces experimental SS curves and that the DeepSDF-based model successfully generates porosity. These results establish a robust pipeline for future large-scale data generation and the reliability-based design of Diecast components.