Cancer is a disease of multicellular animals caused by unregulated cell division. The prevailing model of cancer (multistage carcinogenesis) is based on the view that cancer results after a series of (generally somatic) mutations that knock out the genetic mechanisms suppressing unregulated cell growth. The chance of these mutations occurring increases with size and longevity, leading to Peto's paradox: why don't large animals have a higher lifetime incidence of cancer than small animals? The solution to this paradox is evolution. From an evolutionary perspective, an increasing frequency of prereproductive cancer deaths results in natural selection for enhanced cancer suppression. The expected result is a prereproductive risk of cancer across species that is independent of life history. However, within species, we still expect cancer risk to increase with size and longevity. Here, I review the evolutionary model of cancer suppression and some recent empirical evidence supporting it. Data from humans and domestic dogs confirm the expected intraspecific association between size and cancer risk, while results from interspecific comparisons between rodents provide the best evidence to date of the predicted recruitment of additional cancer suppression mechanisms as species become larger or longer lived.