The physiological equivalents of power output maintenance and recovery during repeated-sprint exercise (RSE) remain to be fully elucidated. In an attempt to improve our understanding of the determinants of RSE performance we therefore aimed to determine its recovery following exhaustive exercise (which affected intramuscular and neural factors) concomitantly with those of intramuscular concentrations of adenosine triphosphate [ATP], phosphocreatine [PCr] and pH values and electromyography (EMG) activity (a proxy for net motor unit activity) changes. Eight young men performed 10, 6-s all-out sprints on a cycle ergometer, interspersed with 30 s of recovery, followed, after 6 min of passive recovery, by five 6-s sprints, again interspersed by 30 s of passive recovery. Biopsies of the vastus lateralis were obtained at rest, immediately after the first 10 sprints and after 6 min of recovery. EMG activity of the vastus lateralis was obtained from surface electrodes throughout exercise. Total work (TW), [ATP], [PCr], pH and EMG amplitude decreased significantly throughout the first ten sprints (P<0.05). After 6 min of recovery, TW during sprint 11 recovered to 86.3±7.7% of sprint 1. ATP and PCr were resynthesized to 92.6±6.0% and 85.3±10.3% of the resting value, respectively, but muscle pH and EMG amplitude remained depressed. PCr resynthesis was correlated with TW done in sprint 11 (r = 0.79, P<0.05) and TW done during sprints 11 to 15 (r = 0.67, P<0.05). There was a ∼2-fold greater decrease in the TW/EMG ratio in the last five sprints (sprint 11 to 15) than in the first five sprints (sprint 1 to 5) resulting in a disproportionate decrease in mechanical power (i.e., TW) in relation to EMG. Thus, we conclude that the inability to produce power output during repeated sprints is mostly mediated by intramuscular fatigue signals probably related with the control of PCr metabolism.