This study was designed to examine the effects of alterations in dietary carbohydrate (CHO) intake on the performance of high-intensity exercise lasting approximately 10 min (EXP 1) and 30 min (EXP 2). Trained subjects exercised to exhaustion on four occasions on a cycle ergometer at 90% of maximal oxygen consumption (VO2max; EXP 1, n = 5) and 80% of VO2max (EXP 2, n = 7). The first two tests were familiarisation trials and were carried out following the subjects' normal diet. Normal training was continued but standardised during the periods of dietary control. The subsequent two tests were performed 2 weeks apart after 7 days of dietary manipulation. The two diets were a 70% and a 40% CHO diet, isoenergetic with each subject's normal diet and administered in a randomised order. At both exercise intensities, time to exhaustion following the high CHO and low CHO diets was not different [mean (SD) EXP 1: 11.56 (3.78) min and 8.95 (2.35) min, P = 0.22; EXP 2: 26.9 (7.4) min and 26.5 (6.5) min, P = 0.90]. No differences in resting blood metabolite concentrations were found apart from a lower beta-hydroxybutyrate (beta-HB) level following the high CHO diet in EXP 2. Blood lactate was higher after exercise at 90% of VO2max following the high CHO diet. Blood lactate was higher, and beta-HB lower during exercise at 80% of VO2max following the high CHO diet. No differences were found in the other blood metabolites tested. The respiratory exchange ratio after 15 min of exercise at 80% of VO2max was higher on the high CHO diet. No differences in oxygen uptake, heart rate (EXP 2) or ratings of perceived exertion (both experiments) were found between conditions. These results indicate that moderate changes in diet composition during training do not affect the performance of high-intensity exercise in trained individuals when the total energy intake is moderately high.