The models measuring the macroeconomic impact of HIV/AIDS are heterogeneous : each one relies on a specific theoretical background. Nevertheless, there are, at least, three main common limits to those approaches : the authors concentrate on the impact on the labour market ; they neglect the potential implications on the capital market ; and they do not model some essential microeconomic impacts such as the change in the agents economic behaviour. More specifically, the analysis of the impact of HIV/AIDS on savings takes into account direct costs such as health expenditures, seldom indirect costs like the anticipation of funeral costs and they do not model differed indirect costs. The paper proposes an analysis of this last kind of implications through the impact of the epidemic on the saving behaviour. This paper focuses on the uncertainty of life expectancy and is based on two frameworks: the Gali (1990) model which considers the life cycle theory with a finite horizon at the aggregate level and the Moresi (1999) model which species a peculiar consumption utility function through uncertain lifetime. The calibration and simulations of our model reveal a significant drop in the future saving rate in South Africa under the hypothesis of a virus evolution similar to the one given by the UN Population Division: the saving rate in 2015, under those hypothesis, should be at least 5 percentage points inferior to the estimated saving rate that would then prevail in the absence of the epidemic.