Nature Energy, Published online: 08 September 2026; doi:10.1038/s41560-026-02134-2

Effective clean technology (cleantech) policies, such as subsidies and carbon pricing, require an understanding of how interventions shape environmental and socioeconomic outcomes. To support targeted, evidence-based policy design, we discuss how causal machine learning can compare alternative policy scenarios and reveal where, how and for whom cleantech policies work.