Plants produce many low molecular mass natural compounds endowed with biological activity. Among them, resveratrol (3,5,4'-trihydroxystilbene) has been demonstrated to be able to affect a plethora of pivotal cellular molecular processes, including transduction pathways and gene expression. These activities result, in turn, in several different cell phenotypes. Particularly, frequent effects of resveratrol treatment appear to be the reduction of growth and the activation of programmed cell death. Accordingly, a number of trials are currently under development to evaluate the possibility of using resveratrol in cancer therapy, both as single agent or in association with other anticancer compounds. However, some reports suggest that, at low concentrations, not only resveratrol does not inhibit the proliferation and/or the survival of cells but, conversely, it induces proliferation and/or protects cells against toxic agents. On the basis of these biphasic effects, it has been proposed that resveratrol belongs to the so-called hormetic compounds. Hormesis is an expression employed by toxicologists to describe a U-shaped (or J-shaped) dose response characterized by a beneficial effect at low doses and a toxic (or inhibitory) activity at high dose. In this review, we will reappraise data that might suggest or disprove that resveratrol is endowed with clear hormetic properties.