Rolling stock planning is one of rail’s hardest operational problems: deciding which physical train runs which service, across a network of constraints around maintenance, distance, and empty kilometres. Working from a real Deutsche Bahn dataset of 190 trips across five German cities, IQM developed and validated a hybrid quantum-classical algorithm that embeds QAOA within a divide-and-conquer framework. The algorithm was executed on IQM Emerald, confirming the full pipeline end-to-end from problem formulation through hardware execution to post-processing. At current circuit depths and subgraph sizes, quantum results are comparable to classical baselines. The framework is in place, understood, and ready to scale.