Introduction. Many investigators have attempted to define the molecular nature of changes responsible for insulin resistance in muscle, but a molecular approach may not consider the overall physiological context of muscle. Because the energetic state of ATP (ΔGATP) could affect the rate of insulin-stimulated, energy-consuming processes, the present study was undertaken to determine whether the thermodynamic state of skeletal muscle can partially explain insulin sensitivity and fuel selection independently of molecular changes. Methods. 31P-MRS was used with glucose clamps, exercise studies, muscle biopsies and proteomics to measure insulin sensitivity, thermodynamic variables, mitochondrial protein content, and aerobic capacity in 16 volunteers. Results. After showing calibrated 31P-MRS measurements conformed to a linear electrical circuit model of muscle nonequilibrium thermodynamics, we used these measurements in multiple stepwise regression against rates of insulin-stimulated glucose disposal and fuel oxidation. Multiple linear regression analyses showed 53% of the variance in insulin sensitivity was explained by 1) VO2max (P = 0.001) and the 2) slope of the relationship of GATP with the rate of oxidative phosphorylation (P = 0.007). This slope represents conductance in the linear model (functional content of mitochondria). Mitochondrial protein content from proteomics was an independent predictor of fractional fat oxidation during mild exercise (R2 = 0.55, P = 0.001). Conclusions. Higher mitochondrial functional content is related to the ability of skeletal muscle to maintain a greater GATP, which may lead to faster rates of insulin-stimulated processes. Mitochondrial protein content per se can explain fractional fat oxidation during mild exercise.