It has long been recognized that endurance training reduces the reliance on carbohydrate as a source of energy during submaximal exercise. Historically, this has been ascribed to a decrease in muscle glycogen utilization. However, recent studies have demonstrated that, at least in humans, training also reduces the production and utilization of plasma-borne glucose during exercise. The latter is true not only during moderate exercise performed at the same absolute intensity before and after training, but also during intense exercise performed at the same relative intensity in the trained and untrained states. Moreover, this adaptation is often quantitatively just as important as the decline in muscle glycogen utilization in accounting for the overall carbohydrate-sparing effect of training. This reduced reliance on plasma glucose, which appears to result from a decrease in muscle glucose transport, seems to be related to the training-induced increase in muscle mitochondrial respiratory capacity. On the other hand, the training-induced decrease in glucose production (which is the result of reductions in both hepatic glycogenolysis and gluconeogenesis) is probably largely due to alterations in the glucoregulatory hormone response to exercise, although other factors (such as changes in hepatic hormone sensitivity and/or responsiveness) may also play a role. By minimizing the possibility of hypoglycemia, these adaptations in glucose production and utilization likely contribute to the increased endurance that results from exercise training.