Immunological tests provide evidence of latent tuberculosis in one third of the global population, which corresponds to more than two billion individuals. Latent tuberculosis is defined by the absence of clinical symptoms but carries a risk of subsequent progression to clinical disease, particularly in the context of co-infection with HIV. In this Review we discuss the biology of latent tuberculosis as part of a broad range of responses that occur following infection with Mycobacterium tuberculosis, which result in the formation of physiologically distinct granulomatous lesions that provide microenvironments with differential ability to support or suppress the persistence of viable bacteria. We then show how this model can be used to develop a rational programme to discover effective drugs for the eradication of M. tuberculosis infection.