What will AI look like in 2030?

Abstract

If scaling persists to 2030, AI investments will reach hundreds of billions of dollars and require gigawatts of power. The report examines AI development trajectories through 2030, arguing that scaling is likely to continue despite substantial challenges—data limitations, electrical power constraints, investment costs, and algorithmic efficiency shifts—each of which appears surmountable based on current trends. By 2030, AI systems are projected to autonomously implement scientific software from natural language descriptions, assist mathematicians in formalizing proofs, and answer complex biology protocol questions. The authors also note that real-world societal impact may substantially lag technical capability development, varying significantly across domains like software engineering versus pharmaceutical development.