Sturgeons are ancient fish exhibiting unique genome plasticity and a high tendency to produce spontaneously autopolyploid genome states. The temperature profiles of the rivers in which sturgeon live and reproduce have been severely altered by human intervention, and the effect of global warming is expected to cause further temperature shifts, which may be detrimental for early developmental stages with narrow windows of thermal tolerance. The comparison of the performance of diploid and autopolyploid sturgeon kept at unfavourable temperatures contributes to scientific knowledge of the effects of polyploid genome states on organisms and can shed light on the ability of polyploids to cope with human‐induced alterations to natural conditions. Using the sterlet Acipenser ruthenus as a model species, we carried out conventional artificial fertilization, as well as the induction of the second polar body retention (SPBR), of the first mitotic division suppression (FMDS) and of the second polar body retention followed by the first mitotic division suppression (SPBR+FMDS). Two experiments were conducted to evaluate the effect of polyploidy on two basic performance parameters, survival and growth. In Experiment 1, fish belonging to untreated, SPBR‐, FMDS‐ and SPBR+FMDS‐induced groups were kept at 10, 16 and 20°C from the neurula stage until the end of endogenous feeding. In Experiment 2, larvae from the untreated and SPBR‐induced groups were reared at 10, 16 and 20°C after their endogenous feeding transition for 3 weeks. Based on our findings, we report that the embryos, prelarvae and larvae of triploid A. ruthenus do not differ from diploids in their ability to survive, grow and develop under suboptimal temperature conditions, while the survival of tetraploids was significantly reduced even at the optimal temperature and even more so at temperatures far from the optimum. This was also the case in the 2n/4n mosaics observed in FMDS‐induced group. Thus, we assume that in tetraploid and 2n/4n individuals, the limits of thermal tolerance are closer to the optimum than in diploids. We also conclude that the hexaploid genome state is probably lethal in A. ruthenus since none of the hexaploids or 3n/6n mosaics arising from the SPBR+FMDS induction survived the prelarval period.