Lactic acid accumulates in contracting muscle and blood beginning at approximately 50-70% of the maximal O2 uptake, well before the aerobic capacity is fully utilized. The classical explanation has been that part of the muscle is O2 deficient and therefore lactate production is increased to provide supplementary anaerobically derived energy. Currently, however, the predominant view is that lactate production during submaximal dynamic exercise is not O2 dependent. In the present review, data and arguments in support of and against the hypothesis of O2 dependency have been scrutinized. Data underlying the conclusion that lactate production during exercise is not O2 dependent were found to be 1) questionable, or 2) interpretable in an alternative manner. Experiments in human and animal muscles under various conditions demonstrated that the redox state of the muscle is reduced (i.e., NADH is increased) either before or in parallel with increases in muscle lactate. Based on experimental data and theoretical considerations, it is concluded that lactate production during submaximal exercise is O2 dependent. The amount of energy provided through the anaerobic processes during steady-state submaximal exercise is, however, low, and the role of lactate formation as an energy source is of minor importance. It is proposed that the achievement of increased aerobic energy formation under conditions of limiting O2 availability requires increases of ADP, Pi, and NADH and that the increases in ADP (and therefore AMP via the adenylate kinase equilibrium) and Pi will stimulate glycolysis, and the resulting increase in cytosolic NADH will shift the lactate dehydrogenase equilibrium toward increased lactate production.