The recent divergence in domestic energy costs between oil and natural gas drives an investigation of the potential for further demand-side utilization of natural gas in the USA. An economic assessment of the US transportation sector was conducted with a focus on the penetration of technologies that use natural gas as a fuel. Bottom-up modeling of the US energy system using the market allocation framework enabled estimation of the optimal technology mix in both the light-duty and heavy-duty vehicle segments, over a 40-year time horizon, under various scenarios of technical learning rates and natural gas prices. A modified functional form of Moore's law was developed and anchored to municipal transit bus data, to represent technical learning with regard to natural gas vehicle costs. Modeling results suggest that the present levels of natural gas vehicle penetration are suboptimal and a number of market failures were identified, which are most likely to propagate under-adoption into the foreseeable future. Some policy guidelines, aimed at the federal level, were outlined as potential responses to the market failures discussed herein. Copyright (c) 2014 John Wiley & Sons, Ltd.