The increasing integration of variable renewable energy sources into power grids introduces significant volatility and uncertainty. This trend challenges system operators to maintain system stability and balance supply and demand using flexible resources. Industrial demand response (DR) represents a promising flexibility option, as industrial facilities can shift or curtail loads to support grid operations. However, uncertainties arising from fluctuating energy demands, renewable generation forecasts, and operational disturbances threaten the reliability and accuracy of industrial DR provision. Stochastic Model Predictive Control (SMPC) is well-suited for managing uncertainty in sequential decision-making problems. Nevertheless, the “curse of dimensionality” associated with large scenario trees often renders real-time SMPC computationally infeasible for industrial applications. As a result, deploying advanced stochastic optimization methods in practice remains difficult. This work addresses the following research question: How can hierarchical energy management architectures incorporate robust SMPC- based DR strategies while overcoming computational limitations and ensuring reliable industrial operation?