The wild relatives of domesticated crops possess genetic diversity useful for developing more productive, nutritious and resilient crop varieties. However, their conservation status and availability for utilization are a concern, and have not been quantified globally. Here, we model the global distribution of 1,076 taxa related to 81 crops, using occurrence information collected from biodiversity, herbarium and gene bank databases. We compare the potential geographic and ecological diversity encompassed in these distributions with that currently accessible in gene banks, as a means to estimate the comprehensiveness of the conservation of genetic diversity. Our results indicate that the diversity of crop wild relatives is poorly represented in gene banks. For 313 (29.1% of total) taxa associated with 63 crops, no germplasm accessions exist, and a further 257 (23.9%) are represented by fewer than ten accessions. Over 70% of taxa are identified as high priority for further collecting in order to improve their representation in gene banks, and over 95% are insufficiently represented in regard to the full range of geographic and ecological variation in their native distributions. The most critical collecting gaps occur in the Mediterranean and the Near East, western and southern Europe, Southeast and East Asia, and South America. We conclude that a systematic effort is needed to improve the conservation and availability of crop wild relatives for use in plant breeding.