A thick layer of fine-grained, ice-rich, aeolian-deposited sediment accumulated in Beringian environments in the late Pleistocene. The terrestrial Arctic permafrost holds an estimated ~1500 Pg C (~40% of total terrestrial soil carbon), being particularly concentrated in the 1.2 million km² of deep soils of the Yedoma regions of Siberia and Alaska (210–460 Pg C) and in Arctic river deltas (91 ± 39 Pg C) (Schuur et al. 2015 Nature 520). The vulnerability of such large quantities of permafrost-stored carbon with rising temperatures and permafrost thaw represents an important climate feedback identified in the high latitudes. Landcover has long been found to affect the energy and carbon budgets of these regions, greatly influencing the pace and direction of such feedback. Central to these mechanisms is the Pleistocene mammoth steppe, which has been proposed to interact with the Earth System differently to the current wet tundra/forest-tundra. Key differences involve higher albedo of grassland-dominated ecosystems – linked to both vegetation reflectance and modified snow cover; reduced surface insulation due to winter snow trampling by large animals; drier soils due to increased evapotranspiration; higher nutrient cycling and productivity through the herbivory-and-egestion pathway; and increased soil C accumulation due to deeper, diffused roots. Independently of how the demise of the Pleistocene megafauna occurred, a key question in Beringian landscapes is whether and to what extent biome-scale vegetation changes were (and might be) controlled bottom-up (climate → vegetation cover → fauna) or top-down (fauna → vegetation cover ↔ climate). The latter opens the possibility for large-scale megafaunal interventions that mitigate climate change by favouring mammoth steppe-like vegetation cover. Whereas much information on which of the two possible mechanisms dominates comes from the palaeo-record, here I will discuss what evidence exists at present – the observational period – from systems where large differences in extant megaherbivore densities are found through experimental, socio-ecological, and/or opportunistic setups in northern Eurasia, including Beringian sites such as Pleistocene Park in Chersky, Sakha. The overall feasibility of such interventions, in case they prove to work, will be assessed.