Nuclear power remains essential to the UK’s energy security and is seen by government as a cost-competitive, low carbon component of the UK’s current and future energy mix, with new nuclear power stations currently under construction along the coast, and more planned to replace aging reactors. The coast offers easy access to cooling waters, but coastal erosion and flooding hazards can pose a significant threat to nuclear facilities and infrastructure, with potentially severe consequences. Implementing effective protection measures over the lifecycle (construction, operation, decommissioning) of these assets is essential and requires robust predictions of shoreline change and flood risk in current and future climates. Nuclear regulations require a Coastal Process Assessment of nearshore wave and water levels, sediment transport and bathymetric change, and predictions of flood depths and coastal erosion/accretion under various climate change projections. Such models are expensive and often carry large uncertainty due to limited datasets and complex interactions between oceanographic and meteorological processes acting at global, regional, and local scales. Hybrid, multivariate statistical-dynamical models are emerging as a skillful tool for probabilistic assessment of coastal hazards and may offer an efficient means for assessing exposure and vulnerability of nuclear sites to climate-mediated coastal hazards.