Limitations in energy supply is a classical hypothesis of muscle fatigue. The present paper reviews the evidence available from human studies that energy deficiency is an important factor in fatigue. The maximal rate of energy expenditure determined in skinned fibres is close to the rate of adenosine triphosphate (ATP) utilisation observed in vivo and data suggest that performance during short bursts of exercise (<5 s duration) primarily is limited by other factors than energy supply (e.g. Vmax of myosine adenosine triphosphatase (ATPase), motor unit recruitment, engaged muscle mass). Within 10 s of exercise maximal power output decreases considerably and coincides with depletion of phosphocreatine. During recovery, maximal force and power output is restored with a similar time course as the resynthesis of phosphocreatine. Increases in muscle store of phosphocreatine through dietary supplementation with creatine increases performance during high-intensity exercise. These findings support the hypothesis that energy supply limits performance during high-intensity exercise. It is well documented that pre-exercise muscle glycogen content is related to performance during moderate intensity exercise. Recent data indicates that the interfibre variation in phosphocreatine is large after prolonged exercise to fatigue and that some fibres are depleted to the same extent as after high-intensity exercise. Despite relatively small decreases in ATP, the products of ATP hydrolysis (Pi and free ADP) may increase considerably. Free ADP calculated from the creatine kinase reaction increases 10-fold both after high-intensity exercise and after prolonged exercise to fatigue. It is suggested that local increases in ADP may reach inhibitory levels for the contraction process.