Preclinical models provided ample evidence that cannabinoids are cytotoxic against cancer cells. Among the best studied phytocannabinoids, cannabidiol (CBD) is most promising for the treatment of cancer as it lacks the psychotomimetic properties of delta-9-tetrahydrocannabinol (THC). In vitro studies and animal experiments point to a concentration- (dose-)dependent anticancer effect. The effectiveness of pure compounds versus extracts is the subject of an ongoing debate. Actual results demonstrate that CBD-rich hemp extracts must be distinguished from THC-rich cannabis preparations. Whereas pure CBD was superior to CBD-rich extracts in most in vitro experiments, the opposite was observed for pure THC and THC-rich extracts, although exceptions were noted. The cytotoxic effects of CBD, THC and extracts seem to depend not only on the nature of cannabinoids and the presence of other phytochemicals but also largely on the nature of cell lines and test conditions. Neither CBD nor THC are universally efficacious in reducing cancer cell viability. The combination of pure cannabinoids may have advantages over single agents, although the optimal ratio seems to depend on the nature of cancer cells; the existence of a ‘one size fits all’ ratio is very unlikely. As cannabinoids interfere with the endocannabinoid system (ECS), a better understanding of the circadian rhythmicity of the ECS, particularly endocannabinoids and receptors, as well as of the rhythmicity of biological processes related to the growth of cancer cells, could enhance the efficacy of a therapy with cannabinoids by optimization of the timing of the administration, as has already been reported for some of the canonical chemotherapeutics. Theoretically, a CBD dose administered at noon could increase the peak of anandamide and therefore the effects triggered by this agent. Despite the abundance of preclinical articles published over the last 2 decades, well-designed controlled clinical trials on CBD in cancer are still missing. The number of observations in cancer patients, paired with the anticancer activity repeatedly reported in preclinical in vitro and in vivo studies warrants serious scientific exploration moving forward.