The trend towards larger offshore wind turbines (WTs) implies the need for bigger support structures. These are commonly derived from existing structures through upscaling and subsequent optimization. To reduce the number of design steps, this work proposes a direct optimization approach, by which means a support structure for a larger WT is obtained through an automated optimization procedure based on a smaller existing system. Due to the suitability of floating platforms for large MW-class WTs, this study is based on the OC3 spar-buoy designed for the NREL 5 MW WT. Using a Python-Modelica framework, developed at Fraunhofer IWES, the spar-buoy geometry is adjusted through iterative optimization steps to finally support a 7.5 MW WT. The optimization procedure focuses on the global system performance in a design-relevant load case. This study shows that larger support structures, appropriate to meet the objective of the hydrodynamic system behavior, can be obtained through automated optimization of existing designs without the intermediate step of upscaling.