Data on growth and composition of dairy calves are available and could be used to better describe nutrient requirements. The objective of this work was to estimate maintenance energy requirements and develop equations to predict energy and protein requirements for growth. Data from 5 comparative harvest studies, 4 with Holstein and one with Jersey calves, representing 206 individuals and 31 different treatments, were compiled and split in 2 sets, for equation development (85%) and evaluation (15%). Average final full body weight (FBW) ranged from 31 to 105 kg. Treatments represented diets with ranges in nutrient supply in protein (14.3–31.2%), fat (14.8–33.4%) and feeding rates (148–353 kcal of metabolizable energy (ME)/d/ kg^0.75 FBW). Data included nutrient intake, weight gain and chemical composition of gain in an empty body weight (EBW) basis. Maintenance energy coefficients were estimated regressing ME intake and the logarithm of the estimated heat production (HP, kcal/d/kg^0.75 EBW). The antilog of the intercept at zero intake was considered the fasting HP, and the value at which the function and the unity line were equal constituted the ME requirement for maintenance (ME_m). Maintenance requirements for metabolizable protein (MP) were calculated using published data. Maintenance requirements were subtracted from nutrient intake to estimate energy (ME_g, Mcal/d/kg^0.75) and protein (MP_g, kg/d/kg^0.75) available for growth. Parameters for predicting retained fat (EBFG, kg/d) and protein (EBPG, kg/d) were estimated using multiple linear regression. Fasting HP was 76 and 85 whereas ME_m was 105 and 129 kcal/kg^0.75 EBW for Holsteins and Jerseys, respectively. The equations to predict EBPG (0.183 × EBW gain (kg/d) + 4.52 x MP_g − 0.344 × ME_g) and EBFG (0.099 × EBW gain (kg/d) + 0.228 × fat intake (kg/d) + 0.152 x ME_g – 8.59 × MP_g + 27.22 × ME_g × MP_g) accounted for 89 and 88% of the variation. Predicted values for EBFP and EBPG were multiplied by their energy constants (9,367 and 5,686 kcal/kg, respectively) and added together to estimate retained energy, which captured 88% of the variation in the evaluated data set.