Abstract This study was conducted to evaluate the effects of dietary NE on growth performance, carcass characteristics, and removal and mortality rates of grow-finish pigs. A total of 1,927 pigs [initial body weight (BW) 23.3 ± 0.23 kg] were used in a 125-d study. Pens of pigs were blocked by BW and randomly allotted to dietary treatments in a randomized complete block design with 23 to 26 pigs/pen and 15 replications/treatment. Pens received 1 of 5 diets with decreasing NE based on Eq. 1-8 NRC (2012) using proximate analysis of the major ingredients. A corn-soybean meal-based control diet was blended with a low energy diet that contained 8% less NE through inclusion of 25% wheat middlings and 15% DDGS (6% oil) to achieve intermediate NE levels. Thus, dietary NE fed through 5 dietary phases were: phase 1 (2,432, 2,383, 2,335, 2,286, and 2,237 kcal/kg); phase 2 (2,463, 2,413, 2,363, 2,315, and 2,266 kcal/kg); phase 3 (2,495, 2,445, 2,395, 2,345, and 2,295); phase 4 (2,518, 2,468, 2,417, 2,367, and 2,317 kcal/kg); and phase 5 (2,537, 2,486, 2,435, 2,384, 2,334 kcal/kg). Dietary phases were fed from approximately 23-40, 40-60, 60-80, 80-105, and 105-133 kg BW, respectively. Overall, reducing NE decreased (linear, P < 0.001) average daily gain (ADG) and final BW and increased (linear, P < 0.001) average daily feed intake (ADFI), resulting in decreased (linear, P < 0.001) gain to feed ratio (G:F; Table). Decreasing NE also tended to increase (linear, P = 0.073) the percentage of removals, with no evidence of difference (P > 0.10) in mortality. Reducing NE worsened (linear, P < 0.001) caloric efficiency (CE) based on NRC nutrient loading values, but improved (linear, P ≤ 0.003) CE based on INRA (2008) or Brazilian Tables for Poultry and Swine (2017) nutrient loading values. When formulating diets using CVB (2021) nutrient loading values, decreasing NE did not influence CE (P > 0.10). For carcass characteristics, decreasing NE decreased (linear, P < 0.001) HCW, backfat depth, and carcass yield. However, decreasing NE increased (linear, P ≤ 0.024) loin depth and percentage lean. When considering CE on a carcass gain basis, decreasing NE worsened (linear, P ≤ 0.004) CE based on NRC or CVB nutrient loading values, but improved (linear, P < 0.001) CE based on INRA nutrient loading values. Decreasing NE resulted in no differences (P > 0.010) in CE based on Brazilian Tables nutrient loading values. These results suggest that the NRC database overestimates NE contributions of fibrous ingredients, such as wheat middlings and corn DDGS, while the INRA database underestimates their contributions to NE, whether CE is calculated based on live or carcass gain. Meanwhile, NE values in the CVB database appear to assess nutrient values more accurately for NE based on live gain but underestimate their value on a carcass basis. The Brazilian tables appear to underestimate the contributions of fibrous ingredients to dietary NE on a live basis, but more accurately value their contributions on a carcass basis.