This study aimed to examine, at the level of the active muscles, whether the plateau in oxygen (O₂) extraction normally observed near the end of a ramp incremental (RI) exercise test to exhaustion, is caused by the achievement of an upper limit in O₂ extraction. 11 healthy men (27.3 ± 3.0 years; 81.6 ± 8.1 kg; 183.9 ± 6.3 cm) performed a RI cycling test to exhaustion. O₂ extraction of the vastus lateralis (VL) was measured continuously throughout the test using the near-infrared spectroscopy (NIRS)-derived deoxygenated hemoglobin [HHb] signal. A leg blood flow occlusion was performed at rest (LBF_OCC 1) and immediately post the RI test (LBF_OCC 2). The [HHb] values during the resting occlusion (108.1 ± 21.7%) (LBF_OCC 1) and the peak values during exercise (100 ± 0%) ([HHb]_PLATEAU) were significantly greater than those observed at baseline (bsln) (0.84 ± 10.6% at bsln 1 and 0 ± 0% at bsln 2) (p < 0.05). No significant difference was found between LBF_OCC 1 and [HHb]_PLATEAU (p > 0.05) nor between the bsln measures (p > 0.05). [HHb] values at LBF_OCC 2 (130.5 ± 19.7%) were significantly greater than all other time points (p < 0.05). These results support the existence of an O₂ extraction reserve in the VL muscle at the end of a RI cycling test and suggest that the observed plateau in the [HHb] signal towards the end of a RI test is not representative of an upper limit in O₂ extraction.