The aim of this study was threefold: (a) to compare the maximal lactate steady state (MLSS) with critical power (CP); (b) to describe the relationship of MLSS with rowing performances; and (c) to verify the agreement of MLSS with several exercise intensity thresholds in rowers. Fourteen male rowers (mean [SD]: age = 26 [13] years; height = 1.82 [0.05] m; body mass = 81.0 [7.6] kg) performed on a rowing ergometer: (I) discontinuous incremental test with 3 min stages and 30-s recovery intervals (INC3min); (II) continuous incremental test with 60-s stages (INC1min); (III) two to four constant workload tests to determine MLSS; and (IV) performance tests of 500, 1000, 2000 and 6000 m to determine CP. Twenty-seven exercise intensity thresholds based on blood lactate, heart rate and ventilatory responses were determined by incremental tests, and then compared with MLSS. CP (257 [38] W) was higher than MLSS (187 [25] W; p < 0.001), with a very large mean difference (37%), large typical error of estimate (14%) and moderate correlation (r = 0.48). Despite the correlations between MLSS and most intensity thresholds (r > 0.70), all presented low correspondence (TEE > 5%), with a lower bias found between MLSS and the first intensity thresholds (-12.5% to 4.1%). MLSS was correlated with mean power during 500 m (r = 0.65), 1000 m (r = 0.86) and 2000 m (r = 0.78). In conclusion, MLSS intensity is substantially lower than CP and presented low agreement with 27 incremental-derived thresholds, questioning their use to estimate MLSS during rowing ergometer exercise.Highlights MLSS was substantially lower than CP in rowing exercise with a mean difference of 37%, much larger than the difference commonly found in running and cycling exercise (i.e., ?10%).A clear disagreement was reported between MLSS and 27 physiological thresholds determined in different incremental tests.There is a positive association of MLSS with 500, 1000 and 2000 m rowing ergometer performance tests.