S06.10 · Energy, Power & Climate
Reducing or shifting customer energy consumption, growing as grid constraints and AI-driven load growth push utilities toward flexibility.
Energy efficiency and demand-side management — reducing or shifting customer energy consumption — is a $30-35B global demand-response market in 2025 (Fortune Business Insights, MarketsandMarkets). The broader building- and industrial-efficiency investment the IEA tracks separately runs to several hundred billion dollars annually but sits mostly outside this segment's relevant scope, and conflating the two flatters the addressable market. Growth runs high-single to low-double digits, driven by grid capacity constraints and AI-driven load growth forcing utilities to seek flexibility. AI's own effect on the segment sits squarely on the cost side, automating load forecasting and program dispatch in a business that is already commercially deployed at scale.
Formal demand-response programs concentrate in the US and EU and are tied to deregulated wholesale markets — a program only pays a customer to shift load when there is a real-time or capacity price to shift against, which is why adoption is weaker in regulated and subsidized-tariff regions where price signals are muted. The segment is fragmented, with aggregators such as Enel X, CPower and Voltus competing alongside utilities' own in-house programs and no dominant global player.
The segment sits between utilities and grid operators on one side and end customers on the other, increasingly integrated with distributed energy resources and storage. Revenue combines contracted capacity payments with performance-based payments, capital intensity is low relative to generation and storage, and regulatory program design is the primary factor gating growth. That low capital intensity is the structural advantage: a megawatt of avoided or shifted demand is typically cheaper to procure than a megawatt of new dispatchable capacity, which is why grid operators facing capacity shortfalls reach for demand response before new build.
Building software and virtual-power-plant integration extend the model outward, converting efficiency programs into dispatchable grid assets rather than simple consumption reductions.
Agentic AI automates load forecasting, building-control optimization, and program enrollment and dispatch, collapsing the manual energy-manager labor cost line; the moat migrates from utility program relationships toward real-time optimization software, and AI-based building energy management is already in broad commercial deployment. The commercial upside for aggregators is volume, not price. Per-megawatt payment rates are set by program design, and no algorithm changes them. What better forecasting and automated dispatch do change is how large a fleet of assets a single aggregator can enroll and manage with the same headcount — and that is what determines an aggregator's share of a given utility program.