Paclitaxel is a successful anti-cancer drug that kills cancer cells in two-dimensional culture through perturbation of mitosis, but whether it causes tumour regression by anti-mitotic actions is controversial. Drug candidates that specifically target mitosis, including inhibitors of kinesin-5, AurkA, AurkB and Plk1, disappointed in the clinic. Current explanations for this discrepancy include pharmacokinetic differences and hypothetical interphase actions of paclitaxel. Here, we discuss post-mitotic micronucleation as a special activity of taxanes that might explain their higher activity in solid tumours. We review data showing that cells which exit mitosis in paclitaxel are highly micronucleated and suffer post-mitotic DNA damage, and that these effects are much stronger for paclitaxel than kinesin-5 inhibitors. We propose that post-mitotic micronucleation promotes inflammatory signalling via cGAS-STING and other pathways. In tumours, this signalling may recruit cytotoxic leucocytes, damage blood vessels and prime T-cell responses, leading to whole-tumour regression. We discuss experiments that are needed to test the micronucleation hypothesis, and its implications for novel anti-mitotic targets and enhancement of taxane-based therapies.