Carbon monoxide (CO) is a ubiquitous atmospheric trace gas produced by natural and anthropogenic sources. Some aerobic bacteria can oxidize atmospheric CO and, collectively, they account for the net loss of ~250 million tonnes of CO from the atmosphere each year. However, the physiological role, genetic basis, and ecological distribution of this process remain incompletely resolved. In this work, we addressed these knowledge gaps using the genetically tractable aerobic actinobacterium Mycobacterium smegmatis. Shotgun proteomic and transcriptional analyses revealed this bacterium upregulates the catalytic subunit of a form I carbon monoxide dehydrogenase by 50-fold when exhausted for organic carbon substrates. Whole-cell biochemical assays in wild-type and mutant backgrounds confirmed that the enzyme oxidizes CO, including at sub-atmospheric concentrations, and supports aerobic respiration. Contrary to current paradigms on CO oxidation, the enzyme did not support chemolithoautotrophic growth and was dispensable for CO detoxification. However, it significantly enhanced long-term survival, suggesting that atmospheric CO serves a supplemental energy source during organic carbon starvation. Phylogenetic analysis indicated that atmospheric CO oxidation is a widespread and ancestral trait of CO dehydrogenases. Homologous enzymes are encoded by 685 sequenced species of bacteria and archaea, and we confirmed genes encoding this enzyme are abundant and expressed in terrestrial and marine environments. CO dehydrogenases are also conserved and expressed in the pathogen Mycobacterium tuberculosis. On this basis, we propose a new survival-centric model for the evolution of CO oxidation and conclude that, like atmospheric H2, atmospheric CO is a major energy source supporting persistence of aerobic heterotrophic bacteria in different environments.