Industrial sector growth in developing countries requires the provision of alternatives to guarantee sustainable development. Improving energy efficiency and fuel switching are two measures to reduce CO2 emissions in the industrial sector, with natural gas and low-carbon electricity as the most feasible options in the short term. In this work, a linear programming optimization model has been developed to study the potential of energy efficiency improvement and fuel substitution for CO2 emissions reduction, at national level in the non-ferrous metals industry. The energy resource/end-use device allocation problem in secondary metal production and semi-fabrication has been modeled. Using this model, the particular case of Colombia, where low-carbon electricity is available, has been studied. By improving energy efficiency, energy use and CO2 emissions can be reduced significantly, 73% and 72%, respectively, at negative costs. Further CO2 emissions reductions, up to 88%, are possible with fuel switching to low-carbon electricity, increasing the costs for the energy system; however, cost reductions caused by energy efficiency improvement outweigh cost increments of fuel switching. Benefits achieved with fuel substitution using low-carbon electricity can be lost if hydropower is not available; in such a case, efficient natural gas-fired end-use devices are preferable. (C) 2012 Elsevier Ltd. All rights reserved.