For manufacturing of large parts made of lightweight materials like aluminum, fiber reinforced plastics or composites for example for the frame in aerospace or automotive industries more and more industrial robots are used. Their main challenge is the low stiffness compared to conventional machine tools resulting in positioning errors. A lot of research is done in order to compensate trajectory errors and enable them for milling operations, which result from the weaknesses in the kinematic. Nevertheless, dynamic properties influence the process stability, which cannot be compensated with the robot control as the dynamic of the joint, and the cycle time of the robot control is limited. Therefore, different robot designs are presented and compared regarding their stiffness, dynamic properties and costs. Afterwards the main weaknesses of the selected design were identified and used for optimization to reduce the deflection and positioning errors during cutting operation. Furthermore, the machine tool structure was topologically optimized for different poses to achieve a higher accuracy in the working space.