The tremendous progress achieved during the last few years with the use of highly active antiretroviral therapy in suppressing HIV replication together with improvements in immunity have been tempered by a growing number of new adverse effects. Mitochondrial toxicity is one aspect of these long-term toxicities of antiretroviral drugs, with the role of nucleoside analogs particularly underlined. Some cases of impaired mitochondrial function have been clearly identified, such as pancreatitis due to didanosine, neuropathy due to zalcitabine, myopathy due to zidovudine, and lactic acidosis due to stavudine. These mitochondrial toxicities can affect several organs, presenting different patterns of symptoms: from asymptomatic to states with few symptoms despite huge metabolic abnormalities whose prognosis is immediately life-threatening. Beyond the inhibition of DNA polymerase gamma using nucleoside analogs, responsible for decreasing mitochondrial DNA in certain targeted organs, it appears that several physiopathologic mechanisms interact to explain this observed toxicity, HIV itself plays a role, and the underlying genetic pool needs to be better identified. Such cases mean that, it is imperative to avoid cumulated toxicities caused by associated treatments. With serious cases, or persistent symptoms despite discontinuing the nucleoside analogs responsible for such toxicity, one must propose vitamins, mitochondrial co-factors, or anti-oxidants. However, the future lies in the use of potent, less toxic nucleoside analogs, and in developing compounds belonging to other classes of antiretrovirals.