The heated debate on automation-vs-labor pulls at the heartstrings of many in both academia and industry, as automation can trigger dramatic adjustments in the urban labor market. Yet, little is known about how migrant workers respond to such technological advances. To contribute to this debate, this paper empirically investigates the effects of automation (proxied by exposure degree of industrial robots) on the settlement intention of migrant workers in urban China. Employing data from the China Migrants Dynamic Survey (CMDS) 2014–2017, our results reveal that cities with greater exposure to industrial robots are more likely to attract migrant workers to settle. Yet, the promotion of settlement intentions is more pronounced for migrant workers at the high and low tails of the skill distribution, leading to a U-shaped selection. In line with our theoretical predictions, mechanism analyses confirm that this nonlinearity is mainly due to the displacement of routine-intensive occupations, the reinstatement of high-skilled machine-complementing occupations, and the indirect booming in service-related occupations at all skill levels. Our findings raise concerns about the polarization of skills and wages among migrant workers despite the fact that they generally benefit from automation.