The Xpert MTB/RIF assay is an accurate test for the diagnosis of tuberculosis when an adequate sputum sample can be obtained; even in smear-negative tuberculosis the sensitivity is about 67%. Although the assay turnaround time is under 2 h, depending on the health-care setting, time to tuberculosis treatment can be 2 weeks or more in a substantial number of patients. The technology has now been endorsed by WHO as a frontline test for tuberculosis in populations where there is a high incidence of HIV. Indeed, several countries in Africa are rolling out Xpert MTB/RIF. However, for expanded and sustained uptake, governments and policy makers require information about the cost-effectiveness of the technology to allow for appropriate planning and allocation of health-care resources. Cost-effectiveness must be balanced against affordability and sustainability. Thus, although the diagnostic accuracy of the technique is not in doubt, questions remain about the cost-effectiveness of the technology given that the overall number of patients treated for tuberculosis can remain unchanged and given the high rates of empirical treatment in resource-poor health-care settings.