Does industrialization leave a lasting legacy after manufacturing declines? I argue that it does. Large manufacturing firms create demand for managerial and technical capabilities that remain valuable as economies transition toward services. I study this mechanism using cross-country evidence, U.S. county data from 1850 to 2000, and a quantitative growth model in which firm scale creates demand for high-skill workers to handle coordination problems, generating organizational knowledge that diffuses across firms and sectors. In the data, countries with larger historical manufacturing peaks develop service sectors more intensive in high-skill producer services and grow faster even after manufacturing has declined. Across counties in the United States, a 10pp higher manufacturing peak employment share driven by national growth in industries where large-scale coordination was technologically important raises the share of high-skill producer services in total private-service employment by 2.8 percentage points in 2000, or 45% of the cross-county average. The calibrated model shows that this mechanism goes a long way toward explaining cross-country differences in service composition, skill accumulation, and growth, and identifies firm scale and the diffusion of organizational knowledge as central forces. Because manufacturing's lasting effects operate through the organizational knowledge created by production at scale rather than manufacturing employment itself, subsidizing manufacturing leaves little persistent legacy. Lasting gains instead arise when skill formation is combined with reforms that allow firms to grow and improve how production is organized.