Lifting hydrofoils are gaining importance, since they drastically reduce the wetted surface area of a ship, thus decreasing resistance. To attain efficient hydrofoils, the geometries can be obtained from an automated optimization process. However, hydrofoil simulations are computationally demanding, since fine meshes are needed to accurately capture the pressure field and the boundary layer on the hydrofoil. Simulation-based optimization can therefore be very expensive. Surrogate models, trained by a limited number of simulations, can reduce the cost for optimization by performing simulations where these are more informative. Furthermore, if an efficient low-fidelity solver is available, multi-fidelity framework can provide a further reduction in required simulations, by combining the accuracy of a few high-fidelity with an exploration of many low-fidelity computations. We propose a hydrofoil optimization procedure based on two simulation codes, a hydrofoil potential flow solver for low-fidelity and a RANS solver for high(er)-fidelity simulations. The RANS solver uses adaptive grid refinement to attain high accuracy with a limited computational budget. Two different multi-fidelity frameworks are compared for a realistic hydrofoil: only RANS based and potential-RANS based. The efficiency for different combinations of fidelity levels is investigated.