AI Power Demand Sets 2026 Record, Shifting ETF Focus to Utilities

Global data center electricity consumption reaches 565 terawatt-hours in 2026. This 26% year-over-year increase stems from AI-optimized server demand. The surge pushes aggregate demand beyond current U.S. economic absorption capacity.

Chicago Fed President Austan Goolsbee warned that this spending creates inflationary pressure. He noted that persistent constraints in electricity and construction could limit growth. This shifts investor focus from software to physical infrastructure providers.

Copper Drives Infrastructure Mineral Demand

A study found power infrastructure drives most modeled mineral demand. Copper accounts for 83% of total mineral mass in AI data centers. This makes copper-focused funds a distinct part of the AI trade.

Investors look to Global X U.S. Infrastructure Development ETF for exposure. It holds companies like Deere and Eaton that build physical networks. These firms supply the heavy equipment and raw materials needed for construction.

Utilities Capture Rising Power Needs

Utilities face record American electricity consumption in 2026. Data center development and manufacturing activity contribute to this rise. The Utilities Select Sector SPDR Fund offers broad exposure to these firms.

Energy sectors provide another route into the theme. Oil and gas represent 91% of the Energy Select Sector SPDR ETF. Exxon Mobil and Chevron account for roughly 35% of its total assets.

Nuclear Offers Specific Infrastructure Exposure

Nuclear power provides a specific angle on AI infrastructure. The Global X Uranium ETF holds 57 securities as of September. Cameco Corp is the largest position at 22.2% of the fund.

Benzinga reports that AI investment may broaden beyond traditional tech. If power and metal constraints persist, infrastructure beneficiaries gain importance. The Fed must decide if this spending forces demand constraints.