Although technological learning is indispensable for economic transformation in developing countries, recent research on industrial policy both lacks consensus regarding policy models and engages in little long-term analysis of policy impacts. This study contributes to this literature through a controlled case comparison of the varied addition of new and unique functional capacities in the Mexican and Brazilian automotive and petroleum industries from 1975 to 2000. It offers a dynamic industrial policy perspective that underscores the explanatory role of alternating state- and market-led industrial policy approaches and their associated cumulative processes of “exploration” and “exploitation” (March (1991)). It also suggests that two background conditions—prior investments in learning and exogenous shocks that undermine the status quo—intervene decisively in the successful sequencing of policy approaches. The study concludes by proposing a framework that recognizes three main learning pathways formed through different configurations of the main independent variable and background conditions. This framework can be deployed as a rough predictive tool to assess how other industries might most effectively increase their technological sophistication.