Human tuberculosis is a complex disease caused by bacterial populations that are located in discrete lesions (microenvironments) in a single host. Some of these microenvironments are conducive to replication, whereas others restrict bacterial growth without necessarily sterilizing the infecting microorganisms. The physical and biochemical milieu in these lesions is poorly defined. None of the existing animal models for tuberculosis (except perhaps non-human primates) reproduce the diversity of disease progression that is seen in humans. Nonetheless, transcriptomics and studies using bacterial mutants have led to testable hypotheses about metabolic functions that are essential for viability in the absence of replication. A complete picture of bacterial metabolism must balance reducing equivalents while maintaining an energized membrane and basic cellular processes.