Sartorius researchers unveiled a nonlinear model predictive control framework intended to improve fully continuous CHO perfusion runs. In reported digital twin experiments, the approach coordinated feed, bleed and harvest flows to increase production by 68% while keeping viable cell density at or above 95%. The control system uses moving estimates for unknown states and a feasibility-aware economic mode to maximize performance in real time. The company says the framework avoids instability problems seen in standard economic NMPC formulations, reducing the need for frequent retuning. For manufacturers, the update targets a core operational bottleneck in continuous processing: maintaining stable biology while optimizing economics under constraints and with accumulating inhibitory biomaterials. If validated in wet runs and scaled across facilities, the method could support more consistent perfusion performance and reduce the engineering burden of continuous intensification.