The increasing complexity of materials discovery necessitates a shift from traditional trial-and-error approaches to fully integrated, AI-driven workflows. In this Perspective, we introduce a constraint-aware, AI-guided framework designed to systematically explore vast chemical and structural spaces and generate novel materials for sustainable energy applications. Our workflow integrates high-throughput, energy-informed computations with machine learning (ML), physics-informed generative models, experimental feedback, and uncertainty quantification, all aligned with sustainability objectives. Central to this approach are advanced ML techniques and generative models that ensure that the proposed materials are both chemically feasible and functionally optimized. We highlight the transformative potential of closed-loop AI-driven discovery to accelerate development across key energy technologies, including batteries, catalysts, photovoltaics, and thermoelectrics. By positioning AI not merely as a predictive tool but as an autonomous research partner, this perspective provides a roadmap for rapidly designing, validating, and deploying next-generation sustainable energy materials.
Fawzy et al. (Thu,) studied this question.