Botulinum neurotoxins (BoNTs) rank amongst the most potent toxins known. The factors responsible for the emergence of the many known and yet unknown BoNT variants remain elusive. It also remains unclear why anaerobic bacteria that are widely distributed in our environment and normally do not pose a threat to humans, produce such deadly toxins. Even the possibility of accidental toxicity to humans has not been excluded. Here, I review the notion that BoNTs may have specifically evolved to target vertebrates. Considering the extremely complex molecular architecture of the toxins, which enables them to reach the bloodstream, to recognize and enter neurons, and to block neurotransmitter release, it seems highly unlikely that BoNT toxicity to vertebrates is a coincidence. The carcass–maggot cycle provides a plausible explanation for a natural role of the toxins: to enable mass reproduction of bacteria, spores, and toxins, using toxin-unaffected invertebrates, such as fly maggots, as the vectors. There is no clear correlation between toxigenicity and a selective advantage of clostridia in their natural habitat. Possibly, non-toxigenic strains profit from carcasses resulting from the action of toxigenic strains. Alternatively, a gene-centered view of toxin evolution would also explain this observation. Toxin-coding mobile genetic elements may have evolved as selfish genes, promoting their own propagation, similar to commensal viruses, using clostridia and other bacteria as the host. Research addressing the role of BoNTs in nature and the origin of toxin variability goes hand in hand with the identification of new toxin variants and the design of improved toxin variants for medical applications. These research directions may also reveal yet unknown natural antidotes against these extremely potent neurotoxins.