Corynebacterium glutamicum is intensively used for the industrial large-scale (fed-) batch production of amino acids, especially glutamate and lysine. However, metabolic flux analyses based on 13C-labeling experiments of this organism have hitherto been restricted to small-scale batch conditions and carbon-limited chemostat cultures, and are therefore of questionable relevance for industrial fermentations. To lever flux analysis to the industrial level, a novel Sensor Reactor approach was developed (El Massaoudi et al., Metab. Eng., submitted), in which a 300-L production reactor and a 1-L Sensor Reactor are run in parallel master/slave modus, thus enabling 13C-based metabolic flux analysis to generate a series of flux maps that document large-scale fermentation courses in detail. We describe the successful combination of this technology with nuclear magnetic resonance (NMR) analysis, metabolite balancing methods and a mathematical description of 13C-isotope labelings resulting in a powerful tool for quantitative pathway analysis during a batch fermentation. As a first application, 13C-based metabolic flux analysis was performed on exponentially growing, lysine-producing C. glutamicum MH20-22B during three phases of a pilot-scale batch fermentation. By studying the growth, (co-) substrate consumption and (by-) product formation, the similarity of the fermentations in production and Sensor Reactor was verified. Applying a generally applicable mathematical model, which included metabolite and carbon labeling balances for the analysis of proteinogenic amino acid 13C-isotopomer labeling data, the in vivo metabolic flux distribution was investigated during subsequent phases of exponential growth. It was shown for the first time that the in vivo reverse C(4)-decarboxylation flux at the anaplerotic node in C. glutamicum significantly decreased (70%) in parallel with threefold increased lysine formation during the investigated subsequent phases of exponential growth.