The Arctic is one of the biomes that undergoes the most dramatic seasonal changes in photoperiod, ambient temperature and primary production. Reduced winter metabolic rate as a response to food scarcity may facilitate survival through the most unfavourable conditions (1). We measured the daily energy expenditure (DEE) of adult female (age 4-8 years) Svalbard reindeer (Rangifer tarandus platyrhynchus) using the doubly labelled water (DLW) technique. Using telemetry data we aimed to quantify adaptations to cold and low food supply in winter by exploring drivers of individual variation in DEE. Individuals were caught in late winter (March-April) in 2017 and 2018 in Nordenskiöld Land, Svalbard, Norway. The animal handling protocol was approved by the Norwegian Food Safety Authority (permit no. 17/237024) and the Governor of Svalbard (permit no. 16/01632-9). Once captured, animals were weighed and dosed with ~0.3 mL of DLW (65 atom% 18O, 35 atom% 2H) per kg of body mass, and re-sampled 2 and 10-20 days later for initial and final isotope concentrations, respectively. All females were fitted with a GPS collar (weight of 0.75 kg) that recorded hourly positions and acceleration in X-Y axes every 5 min. Mean DEE of Svalbard reindeer was 6.2 MJ day-1 (SD = 0.7, n=21). The measured DEE was only 43% of that predicted from allometric scaling for ungulates (2). Mean body mass (BM) was 49.0 kg (SD = 3.4) and mean fat-free mass (FFM; 73% of total body water) was 42.0 kg (SD = 3.7). Using linear regression, we showed that both FFM and activity levels had positive and significant effects on DEE (r2 = 0.51, p < 0.001). FFM and activity levels accounted for 26.6% and 24.2% of the variation in DEE, respectively. We found no effects of displacement by GPS, pregnancy or age on DEE. We then modelled DEE over winter to estimate winter energy budgets, based on activity levels and mean autumn body composition of adult females. Our estimates show that individual winter energy budgets will vary depending on body size, composition (relative amount of fat and FFM) and activity pattern throughout the winter, and that higher relative fat mass (% of BM) determines starvation buffering capacity. Energy reserves contribute only 12% of total winter energy expenditure for a light, lean and active reindeer, while it provides 35% of total winter energy expenditure for a heavy, fat and sedentary reindeer. We demonstrate that Svalbard reindeer expend less energy in the winter than previously assumed, which is likely facilitated by the insulating capacity of their fur and relatively sedentary lifestyle. Our results highlight the importance of individual variation in estimates of energy expenditure in the wild, which may provide a major advance in the assessment of species’ resilience to deteriorating winter conditions in a changing climate (3).