Reducing the uncertainties associated with extrapolation to very long term of corrosion data obtainable from laboratory tests on a relatively young spent nuclear fuel is a formidable challenge. In a geologic repository, spent nuclear fuel may come in contact with water tens or hundreds of thousands of years after repository closure. The corrosion behavior will depend on the fuel properties and on the conditions characterizing the near field surrounding the spent fuel at the time of water contact. This paper summarizes the main conclusions drawn from multiyear experimental campaigns performed at JRC-ITU to study corrosion behavior and radionuclide release from spent light water reactor fuel. The radionuclide release from the central region of a fuel pellet is higher than that from the radial periphery, in spite of the higher burnup and the corresponding structural modifications occurring at the pellet rim during irradiation. Studies on the extent and time boundaries of the radiolytic enhancement of the spent fuel corrosion rate indicate that after tens or hundreds of thousands of years have elapsed, very small or no contribution to the enhanced corrosion rate has to be expected from alpha radiolysis. A beneficial effect inhibiting spent fuel corrosion due to the hydrogen overpressure generated in the near field by iron corrosion is confirmed. The results obtained so far point toward a benign picture describing spent fuel corrosion in a deep geologic repository. More work is ongoing to further reduce uncertainties and to obtain a full description of the expected corrosion behavior of spent fuel.