We tested the theory that links the capacity to perform prolonged exercise with the size of the muscle tricarboxylic acid (TCA) cycle intermediate (TCAI) pool. We hypothesized that endurance training would attenuate the exercise-induced increase in TCAI concentration ([TCAI]); however, the lower [TCAI] would not compromise cycle endurance capacity. Eight men (22 ± 1 yr) cycled at ∼80% of initial peak oxygen uptake before and after 7 wk of training (1 h/day, 5 days/wk). Biopsies (vastus lateralis) were obtained during both trials at rest, after 5 min, and at the point of exhaustion during the pretraining trial (42 ± 6 min). A biopsy was also obtained at the end of exercise during the posttraining trial (91 ± 6 min). In addition to improved performance, training increased ( P < 0.05) peak oxygen uptake and citrate synthase maximal activity. The sum of four measured TCAI was similar between trials at rest but lower after 5 min of exercise posttraining [2.7 ± 0.2 vs. 4.3 ± 0.2 mmol/kg dry wt ( P < 0.05)]. There was a clear dissociation between [TCAI] and endurance capacity because the [TCAI] at the point of exhaustion during the pretraining trial was not different between trials (posttraining: 2.9 ± 0.2 vs. pretraining: 3.5 ± 0.2 mmol/kg dry wt), and yet cycle endurance time more than doubled in the posttraining trial. Training also attenuated the exercise-induced decrease in glutamate concentration (posttraining: 4.5 ± 0.7 vs. pretraining: 7.7 ± 0.6 mmol/kg dry wt) and increase in alanine concentration (posttraining: 3.3 ± 0.2 vs. pretraining: 5.6 ± 0.3 mmol/kg dry wt; P < 0.05), which is consistent with reduced carbon flux through alanine aminotransferase. We conclude that, after aerobic training, cycle endurance capacity is not limited by a decrease in muscle [TCAI].