To investigate the acute effects of hypoxia applied during discrete work and recovery phases of a perceptually regulated, high-intensity interval exercise (HIIE) on external and internal loads in inactive overweight individuals. On separate days, 18 inactive overweight (28.7 ± 3.3 kg m−2; 31 ± 8 years) men and women completed a cycling HIIE protocol (6 × 1 min intervals with 4 min active recovery, maintaining a perceived rating of exertion of 16 and 10 during work and recovery, respectively, on the 6–20 Borg scale) in randomized conditions: normoxia (NN), normobaric hypoxia (inspired O2 fraction ∼0.14) during both work and recovery (HH), hypoxia during recovery (NH) and hypoxia during work only (HN). Markers of external (relative mean power output, MPO) and internal load (blood lactate concentration, heart rate and tissue saturation index (TSI)) were measured. MPO was lower in HH compared to NN, NH and HN (all P textless 0.001), with HN also being lower than NN (P textless 0.001) and NH (P textless 0.023). Heart rate was higher in HN than NN, HH and NH (all P textless 0.001). Blood lactate response was higher in NN than HH (P = 0.003) and NH (P = 0.008). Changes in the TSI area above the curve were greater in HN relative to NN, HH and NH (all P textless 0.001). Hypoxia applied intermittently during the work or recovery phases may mitigate the declines in mechanical output observed when exercise is performed in continuous hypoxia, although hypoxia implemented during the work phase resulted in elevated heart rate and lactate response. Specifically, exercise performance largely comparable to that in normoxia can be achieved when hypoxia is implemented exclusively during recovery.