Our objective was to study how balancing Lys, Met and His (LMH) at 2 supplies of net energy of lactation (NEL) affect milk yield and composition. Four Holstein cows received duodenal infusion of 2 amino acids (AA) mixtures and 2 dietary levels of NEL (E-: 31.5 vs E+: 35.9 Mcal/d), combined in factorial arrangement, distributed in a 4×4 Latin square, with 3-wk periods. The 2 AA mixtures, both providing 92±2 g/d of PDI (metabolisable protein in INRA, 2018), contained (g/d) Lys (28), Met (9), His (9) and non-essential AA (NEAA) in LMH+ or only NEAA in LMH-. Diets plus infusions provided 5.8, 2.0 and 2.0% of PDI of Lys, Met and His respectively in LMH- compared to 7.1, 2.4 and 2.4% PDI in LMH+; LMH+ met INRA (2018) requirements of 7.0, 2.4 and 2.4% PDI for these 3 AA, respectively. The mammary uptake (MU) of metabolites were measured on day 20 (6 samples at 2-h intervals) of each period, mammary plasma flow being estimated using the Fick principle (Phe+Tyr). Analyses of variance of these data included the effects of LMH, NEL and interaction using a MIXED procedure (SAS) with cow as random effect. No interaction was observed on milk composition (P>0.22). Balancing LMH tended to increase (P=0.07) and increasing NEL supply increased (P≤0.02) milk and fat yields averaging respectively 28.2, 29.6, 30.1 and 31.2±1.33 kg/d and 1,289, 1,317, 1,363 and 1,410±55.1 g/d for E-LMH-, E-LMH+, E+LMH- and E+LMH+. Both treatments increased (P<0.05) milk protein yield (828, 916, 927 and 962±39.2 g/d for E-LMH-, E-LMH+, E+LMH- and E+LMH+). The increased milk protein yield were linked to increased MU of LMH in response to LMH supply (+ 8.4 mmol N/h; P=0.01) and to increased MU of essential AA other than LMH in response to NEL supply (+ 11.8 mmol N/h; P=0.05). Increasing NEL supplies increased milk fatty acids de novo synthesised within the mammary gland (+58 mmol C/h; P=0.02) and MU of β-hydroxybutyrate (+171 mmol C/h; P=0.03). It increased milk yield through an increased lactose yield (+ 97 g/d; P=0.01) while LMH supply had no effect on it (P=0.21). Glucose and acetate MU only tended to be higher in E+LMH+ than in the other treatments (LMH × E: P≤0.06). In conclusion, balancing LMH increased milk, protein and fat yields through different mechanisms than increasing NEL supply indicating the high mammary gland metabolic flexibility to nutrient availability.