Uncovering the directionality of innovation in artificial intelligence for energy transitions: A global perspective on the twin transition using patent analysis

The intersection of Artificial Intelligence and energy transition, often referred to as the Twin Transition, is receiving increasing attention among policymakers, technologists and researchers across sectors such as energy, transport, buildings and industry. However, beyond anecdotal evidence, comprehensive understanding of emerging directionality of Twin Transition remains limited. This study uses patent landscape analysis to identify the directionality of Twin Transition, making a conceptual contribution by developing a tentative anticipatory framework spanning geographical, actor, sectoral and functional dimensions, with patent data serving as upstream indicators of innovation orientation. Drawing on the European Patent Office's database (Espacenet) covering the top five jurisdictions (China, the United States, South Korea, Europe and Japan), the analysis shows a high concentration of patent families in China, with the remainder distributed across the other four jurisdictions. Such concentration does not by itself imply technological leadership, and the analysis is positioned as a study of upstream innovation patterns rather than downstream transition outcomes. Corporations dominate patenting activity in Europe, Japan, and the United States, whereas in China and South Korea universities and government/public institutions play a greater role. Artificial Intelligence applications in Information and Communication Technology and for improving efficiency dominate in China and the United States, whereas Japan demonstrates a comparatively stronger industrial orientation. The under-representation of sectors such as buildings constitutes an important directionality finding in itself. Finally, it highlights the uneven and contested nature of Twin Transition and calls for future research to unpack its directionality beyond the Global North, but not limited to China.