The spread of invasive pests is increasing due to greater global transportation and climate change-mediated range shifts. In a new community, these invasive species provide a novel resource that native predators or parasitoids can evolve to utilise. For invasive agricultural pests, this could provide top-down control; however, initial attack rates on invasive species in the wild are generally low. The potential for rapid adaptive evolution of native species to improve developmental success on invasive species is unknown. We address this by focusing on Drosophila suzukii, a polyphagous invasive pest, and two cosmopolitan Drosophila parasitoids found in North America prior to the arrival of D. suzukii (Pachycrepoideus vindemiae and Trichopria drosophilae). We experimentally evolved three replicate populations of both species on D. suzukii and the cosmopolitan D. melanogaster, a common host for both parasitoids. We found evidence that P. vindemiae increased developmental success by 88% (95% credible intervals = [−14%, 254%]) and T. drosophilae increased developmental success by 259% [38%, 711%] on D. suzukii after only three generations of selection, despite very low founding sizes of 2 and 30 wasps, respectively. These results demonstrate that rapid evolution of increased virulence is possible even from low genetic diversity, with implications for integrating evolutionary techniques into biological control of invasive species.