We used 48 animals with initial body weight (BW) of 380.2 ± 4.73 kg to determine protein requirements of Nellore and Angus young bulls by comparative slaughter technique. Four animals of each breed were slaughtered at beginning of the experiment (baseline animals). The remainder were housed in individual stalls, where 8 animals of each breed were fed ad libitum with a whole corn shell (WCS) diet (85% WCS and 15% of a pellet based on soybean meal) or a silage/concentrate (SC) diet (30% silage and 70% concentrate based on corn and soybean meal). Other 4 animals of each breed were fed with the SC diet at 50% of the dry matter intake adjusted for metabolic BW of animals that received the SC diet ad libitum. Intake was measured daily and a metabolism trial was conducted with total collection of feces and urine. These data were used to estimate the crude and metabolizable protein intake (CPI and MPI, respectively). At 90 d of growth the cattle were slaughtered. The net requirement of protein for maintenance (NP_m, g/kg EBW^0.75/d) was assumed to be the intercept of the linear regression of the retained protein (RP) on CPI and the slope of the linear regression of RP on MPI was assumed to be the efficiency of protein utilization for maintenance (k). The NP_m was divided by the k to obtain the metabolizable requirement of protein for maintenance (MP_m, g/kg EBW^0.75/d). The net requirement of protein for growth (NP_g) was calculated by NP_g (g/kg EBG/day) = a + b × RE, where EBG is the empty body gain and RE is the retained energy. The data were analyzed using the GLM procedures of SAS adopting significance level of 0.05. We did not observe an effect (P > 0.05) of breeds on NP_m (1.01 g/kg EBW^0.75/d) and k (38%). MP_m was 2.68 g/kg EBW^0.75/d. In addition, breed did not significantly affect (P > 0.05) on NP_g (NP_g g/day = EBG × [286 – 8.23 × (ER/EBG)]). Funded by FAPEMIG, INCT-CA and CNPq.