Using French administrative data, I provide empirical evidence that larger firms are more likely to adopt automation capital, that adopters grow more in employment and sales relative to non-adopters, and that firms exercise size-based labor market power that increases after adoption. I rationalize this evidence in a general equilibrium model with two key assumptions: automation adoption has scale effects, i.e., automation increases firm productivity, and labor markets are oligopsonistic, finding support for both in the data. The model implies that automation adoption has implications for both equity and efficiency. I find that automation adoption increases median welfare by 3.5%, but it also exacerbates misallocation, leading to efficiency losses. In the absence of automation-induced misallocation, median welfare would have been 4.2% higher, suggesting an important role for efficiency-enhancing policies in conjunction with redistributive policies.