Many unit operations in bioprocessing engineering rely on adequate mixing. During process development, scale-up of unit operations is a critical, often laborious multistep task. Volumetric power input is used as one scale-up criteria for operations in stirred systems. While for stirred tank reactors the volumetric power input can be calculated from the Power Number Ne or extracted from literature, currently microtiter plates (MTPs) require experimental determination of power input for each process condition and geometry. We established data-driven models to predict the volumetric power input in shaken 24-, and 96-well MTPs without further experiments. The models were then applied to predict process conditions for the efficient solubilization of buffer components. Dissolution kinetics obtained in a continuous 10 L bench-scale system with baffles and a Rushton turbine were similar to those determined in 96-well MTPs. Combining the developed model with a geometrical relationship enables the prediction of operating conditions for mixing in a stirred tank reactor up to 10 L from shaken MTPs with 200 µL working volume. This granted a scale-up factor of 50.000, leading to a significant reduction of material and time consumption during process development and makes prediction of conditions more robust compared to the current approaches.