Low aerobic exercise capacity is a risk factor for diabetes and a strong predictor of mortality, yet some individuals are “exercise-resistant” and unable to improve exercise capacity through exercise training. To test the hypothesis that resistance to aerobic exercise training underlies metabolic disease risk, we used selective breeding for 15 generations to develop rat models of low and high aerobic response to training. Before exercise training, rats selected as low and high responders had similar exercise capacities. However, after 8 weeks of treadmill training, low responders failed to improve their exercise capacity, whereas high responders improved by 54%. Remarkably, low responders to aerobic training exhibited pronounced metabolic dysfunction characterized by insulin resistance and increased adiposity, demonstrating that the exercise-resistant phenotype segregates with disease risk. Low responders had impaired exercise-induced angiogenesis in muscle; however, mitochondrial capacity was intact and increased normally with exercise training, demonstrating that mitochondria are not limiting for aerobic adaptation or responsible for metabolic dysfunction in low responders. Low responders had increased stress/inflammatory signaling and altered transforming growth factor-β signaling, characterized by hyperphosphorylation of a novel exercise-regulated phosphorylation site on SMAD2. Using this powerful biological model system, we have discovered key pathways for low exercise training response that may represent novel targets for the treatment of metabolic disease.