The evolution of endothermy represents a major transition in vertebrate history and a major factor underlying the diversity of birds and mammals. Despite the several advantages of an endothermic lifestyle, the tempo and mode of the evolution of endothermy in these lineages remains one of the most controversial subjects in paleontology and evolutionary physiology. Here, we combine a heat transfer model with body size estimates in the theropod phylogeny to reconstruct the evolution of metabolic rates along the bird stem lineage. This model suggests that the continuous reduction in size from ancestral theropod dinosaurs to birds constitutes the evolutionary path of least resistance for endothermy to evolve, because it maximizes thermal niche expansion while obviating the im- pact of elevated energy requirements on population size. In this scenario, metabolic rates would have increased steadily with the accelerated miniaturization observed primarily in the Early-Middle Jurassic (~180 to 170 million years ago), resulting in a rise in metabolic levels along those clades that diverged from the bird stem lineage during this period. Whereas basal theropods would generally exhibit lower metabolic rates, more recent groups such as nonavian maniraptorans were likely decent thermoregulators with higher metabolic rates than similar- sized extant ectotherms. We postulate that the evolution of smaller sizes concomitantly with increased aerobic capacity has shaped the thermal physiology of theropod dinosaurs and driven the emergence of true endothermy within this clade, and we propose a temporal sequence of key evolutionary transitions that might have ultimately resulted in the emergence of small, endothermic, feathered flying dinosaurs. The detailed hypotheses stemming from our model, regarding the variation in metabolic levels across theropod lineages as well as the sequence of events that have led to the radiation of birds, can be readily put to a test and should be subject to further scrutiny.