Coasts contain a disproportionate share of the world’s population, reflecting historical advantages, but environmental change threatens a reversal of coastal fortune in the coming decades as natural disasters intensify and sea levels rise. This paper considers whether large infrastructure investments should continue to favour coastal areas. I use a dynamic spatial equilibrium framework and detailed georeferenced data from Vietnam to examine this issue and find evidence that coastal favouritism has significant costs. Road investments concentrated in coastal regions between 2000 and 2010 had positive returns but would have been outperformed by allocations concentrated further inland even in the absence of sea level rise. Future inundation renders the status quo significantly less efficient. Under a central sea level rise scenario, welfare gains 72% higher could have been achieved by a foresighted allocation avoiding the most vulnerable regions. The results highlight the importance of accounting for the dynamic effects of environmental change in deciding where to allocate infrastructure today.