The aim of this study was to test (a) three methods to estimate the quantity of lactate accumulated (QLaA) in response to supramaximal exercise and (b) correlations between QLaA and the nonoxidative energy supply assessed by the accumulated oxygen deficit (AOD). Nine rowers performed a 3-min all-out test on a rowing ergometer to estimate AOD and lactate accumulation in response to exercise. Peak blood lactate concentration [(La)peak] during recovery was assessed, allowing QLaA(m1) to be estimated by the method of Margaria et al. Application of a bicompartmental model of lactate distribution space to the blood lactate recovery curves allowed estimation of (a) the net amount of lactate released during recovery from the active muscles (NALRmax), and (b) QLaA according to two methods (QLaA(m2) and QLaA(m3)). (La)peak did not correlate with AOD. QLaA(m1), QLaA(m2) and QLaA(m3) correlated with AOD (r = 0.70, r = 0.85 and r = 0.92, respectively). These results confirm that (La)peak does not provide reliable information on nonoxidative energy supply during supramaximal exercise. The correlations between AOD and QLaA(m2) and QLaA(m3) support the concept of studying blood lactate recovery curves to estimate lactate accumulation and thus the contribution of nonoxidative pathway to energy supply during supramaximal exercise.