Retrotransposons had an important role in genome evolution, including the formation of new genes and promoters and the rewiring of gene networks. However, it is unclear how such a repertoire of functions emerged from a relatively limited number of source sequences. Here we show that DNA editing, an antiviral mechanism, accelerated the evolution of mammalian genomes by large-scale modification of their retrotransposon sequences. We find numerous pairs of retrotransposons containing long clusters of G-to-A mutations that cannot be attributed to random mutagenesis. These clusters, which we find across different mammalian genomes and retrotransposon families, are the hallmark of APOBEC3 activity, a potent antiretroviral protein family with cytidine deamination function. As DNA editing simultaneously generates a large number of mutations, each affected element begins its evolutionary trajectory from a unique starting point, thereby increasing the probability of developing a novel function. Our findings thus suggest a potential mechanism for retrotransposon domestication.