Agriculture expansion is already the primary cause of terrestrial biodiversity loss globally1,2; yet, to meet the demands of growing human populations, production is expected to have to double by 2050.3 The challenge of achieving expansion without further detriment to the environment and biodiversity is huge and potentially compounded by climate change, which may necessitate shifting agriculture zones poleward to regions with more suitable climates,4 threatening species or areas of conservation priority.5,6,7 However, the possible future overlap between agricultural suitability and wilderness areas, increasingly recognized for significant biodiversity, cultural, and climate regulation values, has not yet been examined. Here, using high-resolution climate data, we model global present and future climate suitability for 1,708 crop varieties. We project, over the next 40 years, that 2.7 million km2 of land within wilderness will become newly suitable for agriculture, equivalent to 7% of the total wilderness area outside Antarctica. The increase in potentially cultivable land in wilderness areas is particularly acute at higher latitudes in the northern hemisphere, where 76.3% of newly suitable land is currently wilderness, equivalent to 10.2% of the total wilderness area. Our results highlight an important and previously unidentified possible consequence of the disproportionate warming known to be occurring in high northern latitudes. Because we find that, globally, 72.0% of currently cultivable land is predicted to experience a net loss in total crop diversity, agricultural expansion is a major emerging threat to wilderness. Without protection, the vital integrity of these valuable areas could be irreversibly lost.