Robust design of forming processes is gaining attention throughout the industry. Many numerical robustoptimization approaches have been proposed in literature and applied to a range of metal forming appli-cations, often making use of a metamodel-based approach. However, published research is confinedto numerical studies only. Experimental validation of the numerically predicted and improved processrobustness is lacking. In this work, a deterministic and a robust optimization study is performed consid-ering a stretch-drawing process of a hemispherical cup, covering both numerical work and experimentalvalidation. For the robust optimization study, both the effect of material property scatter and processscatter are taken into account. For quantifying the material scatter that can be encountered in a produc-tion environment, samples from 37 coils of a drawing quality forming steel DX54D+Z (EN 10327:2004)from multiple casts have been collected. The numerically determined deterministic and robust optimumis subsequently validated by 2 sets of experiments using the collective of materials. The predicted dif-ference in robustness between both optima, and the improved robustness of the robust optimum is alsoobserved in the experiments. This demonstrates how robust optimization can assist in further stretchingthe limits of forming processes, while remaining robust with respect to sources of variation.