One-third of the world population is infected with Mycobacterium tuberculosis (MTB) and hence at risk of developing active tuberculosis (TB). Each year, 8.8 million patients are newly diagnosed with active TB and 1.6 million patients die of TB. The rapid spread of the human immunodeficiency virus (HIV) has fueled the TB epidemic, especially in sub-Saharan Africa, where 28% of TB patients are HIV positive. The current first-line treatment for TB is a multidrug regimen consisting of rifampin, isoniazid, pyrazinamide, and ethambutol (RHZE), which must be taken for at least 6 months to achieve high cure rates (more than 95% in experimental settings). There are several major problems associated with the currently available TB treatment, including nonadherence due to long and complex treatment leading to suboptimal response, drug resistance, and disease spread; common adverse events contributing to nonadherence; increasing incidence of multidrug-resistant (MDR; resistance to at least rifampin and isoniazid) and extensively drug-resistant (XDR; MDR plus resistance to a fluoroquinolone and an aminoglycoside) TB; challenges in combined TB-HIV treatment such as high pill count, overlapping toxicity, drug interactions, and immune reconstitution syndrome; nonadherence in latent TB prophylaxis with isoniazid; and limitations of the WHO's DOTS strategy (labor-intensive, expensive, suboptimal TB detection via sputum microscopy). Clearly, there is an urgent need to improve treatment by either enhancing existing agents or introducing new drugs. Potential new agents should reduce treatment duration, have an acceptable tolerability profile, be active against MDR/XDR TB, be useful in HIV-infected patients with TB, and be active against latent TB. The aim of this article is to review the challenges of developing new anti-TB drugs, present an up-to-date and critical evaluation of new agents in clinical testing, and suggest ways forward to improve TB treatment.