During maximal hypoxic exercise, a reduction in cerebral oxygen delivery may constitute a signal to the central nervous system to terminate exercise. We investigated whether the rate of increase in frontal cerebral cortex oxygen delivery is limited in hypoxic compared to normoxic exercise. We assessed frontal cerebral cortex blood flow using near-infrared spectroscopy and the light-absorbing tracer indocyanine green dye, as well as frontal cortex oxygen saturation (S(tO2)%) in 11 trained cyclists during graded incremental exercise to the limit of tolerance (maximal work rate, WRmax) in normoxia and acute hypoxia (inspired O2 fraction (F(IO2)), 0.12). In normoxia, frontal cortex blood flow and oxygen delivery increased (P < 0.05) from baseline to sub-maximal exercise, reaching peak values at near-maximal exercise (80% WRmax: 287 ± 9 W; 81 ± 23% and 75 ± 22% increase relative to baseline, respectively), both leveling off thereafter up to WRmax (382 ± 10 W). Frontal cortex S(tO2)% did not change from baseline (66 ± 3%) throughout graded exercise. During hypoxic exercise, frontal cortex blood flow increased (P = 0.016) from baseline to sub-maximal exercise, peaking at 80% WRmax (213 ± 6 W; 60 ± 15% relative increase) before declining towards baseline at WRmax (289 ± 5 W). Despite this, frontal cortex oxygen delivery remained unchanged from baseline throughout graded exercise, being at WRmax lower than at comparable loads (287 ± 9 W) in normoxia (by 58 ± 12%; P = 0.01). Frontal cortex S(tO2)% fell from baseline (58 ± 2%) on light and moderate exercise in parallel with arterial oxygen saturation, but then remained unchanged to exhaustion (47 ± 1%). Thus, during maximal, but not light to moderate, exercise frontal cortex oxygen delivery is limited in hypoxia compared to normoxia. This limitation could potentially constitute the signal to limit maximal exercise capacity in hypoxia.