Artificial intelligence-assisted development of high-energy lithium metal batteries

High-energy lithium metal batteries are considered promising power sources for emerging applications such as electric vehicles, green ships, and unmanned aerial vehicles, owing to their potential to deliver energy densities beyond conventional lithium-ion batteries. Nevertheless, their practical development is seriously restricted by complex electrochemical challenges, including dynamic interfacial evolution under working conditions, cross-scale correlations among composition, structure, transport, mechanics, and electrochemical performance, as well as intertwined reaction mechanisms that are difficult to decouple using conventional experimental or theoretical approaches. Artificial intelligence (AI) provides new opportunities for data acquisition, feature extraction, mechanism analysis, performance prediction, and materials optimization. However, battery-oriented AI should go beyond conventional data-driven models by deeply integrating electrochemical knowledge, physical principles, and interpretable descriptors. In this report, an electrochemistry-driven AI framework is introduced for mechanism mining and materials discovery in high-energy lithium metal batteries. First, battery knowledge is incorporated into feature design to construct interpretable descriptors for key battery components, including sulfur cathode catalysts, electrode architectures, lithium metal anodes, and electrolyte-derived interphases. Representative examples include binary descriptors coupling electronic and structural effects for sulfur redox catalysis, electrode structure design factors correlating mass transport and charge transfer, and solid electrolyte interphase omics descriptors linking interphase composition with lithium deposition behavior. Building on these descriptors, machine learning is further employed to uncover hidden relationships in complex electrochemical systems, decouple electronic, structural, ensemble, and interfacial effects, and quantify their individual contributions to battery performance. Finally, the obtained mechanistic insights are fed back into materials design, enabling the discovery of high-potential sulfur catalysts, regulation of local chemical environments, and molecular skeleton programming of sulfur electrochemistry premediators. These strategies guide the construction of Ah-level lithium–sulfur pouch cells with high energy density and long-term stability.