The Indian subcontinent faces a population increase from 1.6 billion in 2000 towards 2 billion around 2050. Therefore, expansion of agricultural area combined with increases in productivity will be necessary to produce the food needed in the future. However, with pressure on water resources already being high, and potential effects of climate change still uncertain, the question rises whether there will be enough water resources available to sustain this production. The objective of this study is to make a spatially explicit quantitative analysis of water requirements and availability for current and future food production in five South Asian basins (Indus, Ganges, Brahmaputra, Godavari and Krishna), in the absence or presence of two different adaptation strategies: an overall improvement in irrigation efficiency, and an increase of reservoir storage capacity. The analysis is performed by using the coupled hydrology and crop production model LPJmL. It is found that the Godavari and Krishna basins will benefit most from an increased storage capacity, whereas in the Ganges and the Indus water scarcity mainly takes place in areas where this additional storage would not provide additional utility. Increasing the irrigation efficiency will be beneficial in all basins, but most in the Indus and Ganges, as it decreases the pressure on groundwater resources and decreases the fraction of food production that would become at risk because of water shortage. A combination of both options seems to be the best strategy in all basins. The large-scale model used in this study is suitable to identify hotspot areas and support the first step in the policy process, but the final design and implementation of adaptation options requires supporting studies at finer scales.