ABSTRACT: Interaction with non-rigid objects poses a challenge for manufacturing systems. The dimensions, locations and physical properties of soft-tissue objects like fruit or meat are subject to considerable variation and uncertainty. Possible ways to accommodate this are to use a number of different types of sensors and actors. However, a control system physically located next to a manufacturing system is often limited in terms of computational power and flexibility. New or different sensors are difficult to integrate, and the processing of their data cannot be accomplished in a feasible time. In the wake of Industry 4.0, cloud-based control emerged, which may provide a solution for the above-mentioned situations. A cloud environment provides hardware and software resources that allow for fast processing. The support of different programming languages and frameworks reduces development effort. This paper examines challenges for robots in the two application areas of high-value crop harvesting and meat processing and derives requirements for control systems interacting with soft-tissue.