Breaking in or breaking through? How local specialisations shape the integration of AI technologies

Amid global investments in Artificial Intelligence (AI), understanding how this technology is successfully integrated into national innovation systems is a critical challenge. We analyse global patent data (1974–2018) for the world’s leading AI economies (China, Japan, South Korea, and the USA), introducing a novel method to distinguish between two national integration pathways: ‘break-in’ trajectories, where AI is anchored to a country’s existing knowledge base, and ‘break-through’ trajectories, where countries follow global technology patterns independently of domestic knowledge bases. Our findings reveal that durable AI leadership emerges overwhelmingly from ‘break-in’ patterns, where AI specialisations persist by aligning with broader national capabilities. This suggests that the dominant integration mechanism is strategic anchoring: a process where a new technology is adapted and hybridised with existing industrial structures. We argue that AI acts as a ‘recombinant bridge’, enabling this hybridisation and creating multiple entry points for national deployment. These findings imply that effective AI policy requires a ‘smart specialisation’-like approach. The goal should be to foster a symbiotic ecosystem that fuses core AI skills with unique industrial strengths through strategic anchoring. This, however, may lead to the rise of ‘hybrid giants’ that dominate multiple sectors, presenting a new frontier of regulatory and geopolitical challenges.