Former US energy secretary Steven Chu once called energy efficiency “fruit lying on the ground,” but adoption lags due to complexity. AI now identifies inefficiencies in factories and buildings, analyzing vast data from sensors and digital twins. “It’s beyond the power of anyone” to handle such volumes manually, says Brian Motherway of the IEA. In agriculture, AI optimizes irrigation and greenhouse heating, reducing energy use. However, AI’s own energy consumption, needing significant computing power, complicates decarbonization efforts. “AI must do heavy lifting” to offset its footprint, warns Sam Kimmins.

Barriers persist: CEOs prioritize visible renewables over incremental efficiency gains, like “boiler upgrades” over solar panels. Upgrading equipment factory-by-factory is costly and bespoke, admits César Quilodrán-Casas. Yet AI paired with digital twins—simulating facilities virtually—accelerates adjustments. A study found AI-powered models cut unplanned downtime by 35%, boosted energy output by 8.5%, and reduced costs by 26% in renewables. Similar gains are feasible in agriculture, where AI predicts watering patterns to minimize energy.

While AI aids decision-making, experts stress it’s not a standalone solution. “Systems must be ready” for AI, says Stephen Horrax of Ramboll, emphasizing infrastructure and policy. The Climate Group’s Sam Kimmins notes enthusiasm for AI could spur investment, as simple solutions—like $1M efficiency retrofits—save billions. “We need both tech and will” to act, Motherway concludes. The debate pivots on balancing AI’s potential against its energy costs, with industry seen as the linchpin for progress.

AI-driven efficiency gains could reshape decarbonization, but require systemic readiness. As Kimmins quips, “Excitement” around AI might finally push stalled upgrades. The critical question remains: Can AI’s data prowess outpace its own hunger for power? At stake are $2.1B in potential annual savings if adoption scales—a figure underscoring the urgency of overcoming inertia.