1.Dispersal is a key mechanism enabling species to adjust their geographic range to rapid global change. However, dispersal is costly and environmental modifications are likely to modify the cost-benefit balance of individual dispersal decisions, for example, by decreasing functional connectivity. 2.Dispersal costs occur during departure, transience and settlement, and are levied in terms of energy, risk, time and lost opportunity, potentially influencing individual fitness. However, to the best of our knowledge, no study has yet quantified the energetic costs of dispersal across the dispersal period by comparing dispersing and philopatric individuals in the wild. 3.Here, we employed animal-born bio-loggers on a relatively large sample (N = 105) of juvenile roe deer to estimate energy expenditure indexed using the vector of dynamic body acceleration (VeDBA) and mobility (distance travelled) in an intensively monitored population in the south west of France. We predicted that energy expenditure would be higher in dispersers compared to philopatric individuals. We expected costs to be i/ particularly high during transience, ii/ especially high in the more fragmented areas of the landscape, and iii/ concentrated during the night to avoid disturbance caused by human activity. 4.There were no differences in energy expenditure between dispersers and philopatric individuals during the pre-dispersal phase. However, dispersers expended around 22% more energy and travelled around 63% further per day than philopatric individuals during transience. Differences in energy expenditure were much less pronounced during the settlement phase. The costs of transience were almost uniquely confined to the dawn period, when dispersers spent 23% more energy and travelled 112% further than philopatric individuals. Finally, the energetic costs of transience per unit time and the total distance travelled to locate a suitable settlement range were higher in areas of high road density. 5.Our results provide strong support for the hypothesis that natal dispersal is energetically costly and indicate that transience is the most costly part of the process, particularly in fragmented landscapes. Further work is required to link dispersal costs with fitness components so as to understand the likely outcome of further environmental modifications on the evolution of dispersal behaviour. This article is protected by copyright. All rights reserved.