During the 1990s the epidemic's character changed in most African regions that it had first infected. HIV prevalence grew more slowly, stabilised, or began to decline, either because deaths overtook the incidence of new infections or, more controversially, because incidence also declined. Preventive measures claimed their first successes, provoking debate on which measures were effective and how they might be replicated elsewhere. Instead of spreading rapidly among people with high-risk behaviour, the virus increasingly attacked vulnerable members of the general population, especially poor women. In short, the epidemic matured. It remained widespread and fatal, but without the explosive quality of its youth, although with the equally alarming prospect of permanence. This, however, was true only in areas of early infection. Even in the mid 2000s there were still parts of the continent, hitherto isolated by distance, disorder, or some other obstacle, where HIV was spreading with epidemic vigour. At the end of 2004, an estimated 25.4 million people in sub-Saharan Africa had HIV, one million more than two years earlier.

The first half of the 1990s was the worst period in the epidemic's early history. Expansion in southern Africa was at its peak. Preventive measures seemed to make no impact on behaviour. There was no progress with vaccines or curative drugs. Many national programmes were corrupt or ineffective. International interest was low and aid was declining. The WHO's Global Programme, attacked from within and without, was wound up at the end of 1995, leaving many national programmes bereft of funds.

News from Uganda restored hope. In retrospect, epidemiologists came to believe that the incidence of new infections in southern Uganda had peaked in the later 1980s and then fallen rapidly. The first indication that HIV prevalence there was declining came from the antenatal clinic at Mulago Hospital in Kampala, where infection among pregnant women fell between 1989 and 1993 from 28.1 to 16.2 per cent. Other urban hospital clinics in southern Uganda showed similar reductions from 1991-3, as did rural Rakai district. The Aids Control Programme checked its testing procedures before giving the news maximum publicity during 1995.