We examined the effects of increasing hypoxia severity on repeated-sprint running performance and neuromuscular fatigue. Thirteen active males completed eight sprints of 5 s (recovery = 25 s) on a motorized sprint treadmill in normoxia (sea level, SL; F I,O 2 = 0.21), in moderate hypoxia (MH; F I,O 2 = 0.17) and in severe hypoxia (SH; F I,O 2 = 0.13). After 6 min of passive recovery, in all conditions a second set of four sprints of 5 s was conducted in normoxia. Neuromuscular function of the knee extensors was assessed at baseline (Pre-) and 1 min after set 1 (Post-set 1) and set 2 (Post-set 2). In set 1, the mean distance covered in SL (22.9 ± 1.2 m) was not different to MH (22.7 ± 1.3 m; P = 0.71) but was greater than in SH (22.3 ± 1.3 m; P = 0.04). No significant differences between conditions for mean distance occurred in set 2. There was a decrease in maximal voluntary contraction torque (Δ = −31.4 ± 18.0 N m, P < 0.001) and voluntary activation (%VA; Δ = −7.1 ± 5.1%, P = 0.001) from Pre-to Post-set 1, but there was no effect of hypoxia. No further change from Post-set 1 to Post-set 2 occurred for either maximal voluntary contraction or %VA. The decrease in potentiated twitch torque in SL (Δ = −13.3 ± 5.2 N m) was not different to MH (Δ = −13.3 ± 6.3 N m) but was lower than in SH (Δ = −16.1 ± 4 N m) from Pre-to Post-set 1 (interaction, P < 0.003). Increasing severity of normobaric hypoxia, up to an equivalent elevation of 3600 m, can increase indices of peripheral fatigue but does not impact central fatigue after 'all-out' repeated-sprint running.