DFT is generally good for well characterized phenomenon. Superconductivity is a bit of an edge case, really dependent on quantum effects that you need ab-initio quantum mechanics simulations to predict.