<p>The structural integrity and safety of advanced high-strength steel (AHSS) components, particularly in the manufacturing of battery electric vehicles (BEVs), depend critically on weld quality. Conventional weld evaluation techniques are often costly and time-intensive, relying heavily on destructive testing and iterative process adjustments. To overcome these limitations, this study introduces a hybrid AI-digital twin framework (DT Core) for virtual assessment of weld quality in high-strength resistance spot welding (RSW). In this study, Simufact weld simulations are integrated with a multidisciplinary design optimization (MDO) platform (modeFRONTIER) to enable virtual optimization and scheduling of welding parameters. A combined design of experiments (DoE) and finite element method (FEM) strategy, along with trained response surface methodology (RSM) models, was applied to predict weld nugget geometry and reduce process variability. In the automotive industry, comprehensive characterization of AHSS materials provided the basis for accurate DT calibration and validation, while machine learning (ML) algorithms further enhanced predictive performance, substantially reducing reliance on physical trials. The resulting DT achieved high fidelity, with simulation results deviating less than 4% from experimental data. Here, a strong correlation between simulated and measured weld outcomes confirms the model’s accuracy. This work delivers a scalable virtual-to-physical (V2P) solution for robust and cost-efficient virtual weld quality evaluation, advancing precision manufacturing and accelerating process innovation within Industry 4.0 frameworks.</p>