Data Centers Drove Half of U.S. Power Demand Growth in 2025, IEA Says

Global electricity demand rose 3% in 2025, nearly triple the 1.3% increase in overall energy demand, driven by data centers, electric vehicles, and heating needs. U.S. electricity demand grew 2%, with the buildings sector accounting for 80% of the increase and data centers contributing about half of the rise. Solar PV supplied over a quarter of global energy demand growth, ahead of natural gas, underscoring continued momentum in the energy transition.

Analysis

The key market implication is that the load-growth mix is shifting from cyclical to structural: electricity intensity is rising because digital infrastructure and electrification are becoming the marginal driver of demand, while legacy energy demand is only barely growing. That matters because utilities, grid equipment, gas peakers, and power infrastructure owners get a longer runway, whereas pure renewable developers may not capture the full benefit unless they own the bottlenecks—interconnection, transformers, storage, and transmission.

The second-order winner is not “power generation” broadly, but anything that monetizes grid scarcity and reliability. AI/data-center load growth forces utility capex up, which can support regulated asset growth and rate-base expansion, but it also raises execution risk: permitting delays, transformer shortages, and higher financing costs can create a mismatch between load additions and delivered earnings over the next 12-24 months. That tends to favor incumbents with existing interconnects and power contracts over smaller developers chasing greenfield projects.

A contrarian read: the market may be overpricing a straight-line bullish case for solar as a demand beneficiary. If solar is absorbing a large share of incremental demand, near-term upside is constrained by intermittency and the need for firming, which shifts economics toward gas, storage, and transmission rather than panels alone. The reversal risk is on the demand side only if data-center builds slip or if AI capex pauses; otherwise the cleaner trade is to own the grid bottleneck, not the commodity electrons.

Key Decisions for Investors

• Long NEE / XLU into 6-12 months: regulated utilities with AI-load exposure should see rate-base growth reaccelerate; use a 5-8% downside stop because higher rates or regulatory pushback can compress multiples.
• Long HUBB or ETN over pure-play solar installers for 6-12 months: grid equipment and electrification beneficiaries capture the capex cycle with better pricing power; pair against TAN if seeking a relative-value expression.
• Long VST / CEG on any 3-6 month pullback: merchant and low-carbon firm power names should benefit from rising capacity value as intermittent supply becomes less reliable; risk is policy intervention or power-price normalization.
• Short small-cap solar developers or installers versus long transmission names (PWR, MYRG) over 6-18 months: transmission and interconnect buildout is the real bottleneck, while project-level solar economics face margin compression and schedule risk.
• Optionality idea: buy 6-12 month call spreads on utility/gas infrastructure exposure and finance partly by selling upside in high-multiple clean-tech names; this expresses the view that the market underestimates firm-power and grid-spend duration.