Molecular generation and its applications in energetic molecule design: Advances, opportunities, and challenges

Energetic molecules (EMs) are strategically important for national defense, aerospace engineering, and industrial blasting. Generative artificial intelligence (AI) enables efficient chemical space exploration and precise inverse molecular design, accelerating the rational, intelligent development of high-performance EMs. However, AI-driven EM design remains challenging due to the need to efficiently explore the vast structure–property space while generating candidates with high energy density, enhanced safety, and synthetic feasibility. This review covers the fundamentals of molecular generation and recent advances, including key databases, molecular representations, generative architectures, and evaluation metrics. It traces technological progress, classifies and analyzes latest models by representation method, and highlights emerging trends. Representative studies on EM generation are reviewed, key technical bottlenecks are analyzed, and corresponding solutions are proposed to address these challenges. Furthermore, a streamlined closed-loop research framework for the intelligent generation of EMs is proposed, future development directions are outlined, and strategic insights are provided to guide subsequent design efforts.