Reducing the rate of global biodiversity loss is a major challenge facing humanity(1), as the consequences of biological annihilation would be irreversible for humankind(2-4). Although the ongoing degradation of ecosystems(5,6) and the extinction of species that comprise them(7,8) are now well-documented, little is known about the role that remaining wilderness areas have in mitigating the global biodiversity crisis. Here we model the persistence probability of biodiversity, combining habitat condition with spatial variation in species composition, to show that retaining these remaining wilderness areas is essential for the international conservation agenda. Wilderness areas act as a buffer against species loss, as the extinction risk for species within wilderness communities is-on average-less than half that of species in non-wilderness communities. Although all wilderness areas have an intrinsic conservation value(9,10), we identify the areas on every continent that make the highest relative contribution to the persistence of biodiversity. Alarmingly, these areas-in which habitat loss would have a more-marked effect on biodiversity-are poorly protected. Given globally high rates of wilderness loss(10), these areas urgently require targeted protection to ensure the long-term persistence of biodiversity, alongside efforts to protect and restore more-degraded environments.