The Fed Is Watching AI Inflation. Here's What ETF Investors Should Watch

ETF investors have an unusual second-order trade due to exorbitant AI spending.

Chicago Fed President Austan Goolsbee warned Monday that a surge in AI investment is pushing aggregate demand beyond what the U.S. economy can comfortably absorb.

If AI spending keeps running into constraints around electricity, construction capacity, equipment and raw materials, the companies supplying those inputs could gain increasing importance, even as persistent inflation keeps pressure on the broader market.

AI’s Power Problem

Gartner expects global data center electricity consumption to reach 565 terawatt-hours in 2026, up 26% year over year, with AI-optimized servers accounting for 31% of data-center power consumption.

The U.S. Energy Information Administration also expects American electricity consumption to hit a record in 2026, with data-center development and manufacturing activity contributing to the increase.

That makes utilities and power infrastructure a potential second-order AI trade.

The Utilities Select Sector SPDR Fund (NYSE:XLU) provides broad exposure to U.S. utilities, while the Global X U.S. Infrastructure Development ETF (BATS:PAVE) owns companies involved in raw materials, heavy equipment, engineering and construction.

The Raw-Material Bottleneck

A recent study of AI data-center infrastructure found that power infrastructure, rather than semiconductors, drives most modeled mineral demand, with copper accounting for 83% of the total modeled mineral mass.

That puts copper-focused ETFs such as Global X Copper Miners ETF (NYSE:COPX) on a different part of the AI trade.

The investment case is not simply that AI companies need more computing power. It is that building the physical infrastructure required for that computing power requires metals, transmission equipment and construction capacity.

PAVE’s holdings illustrate that exposure. Its portfolio includes equipment and infrastructure companies such as Deere & Co (NYSE:DE), Fastenal Co (NASDAQ:FAST), Nucor Corp (NYSE:NUE), Emerson Electric Co (NYSE:EMR), Eaton Corporation PLC (NYSE:ETN) and Quanta Services Inc (NYSE:PWR).

Energy Is Another Link

Energy could provide another route into the theme, particularly if persistent oil and commodity shocks remain part of the inflation equation.

The Energy Select Sector SPDR ETF (NYSE:XLE) is heavily concentrated in oil and gas, with oil, gas and consumable fuels representing about 91% of the fund as of Sept. 17. Exxon Mobil Holdings Corp (NYSE:XOM) and Chevron Corporation (NYSE:CVX) together accounted for roughly 35% of assets.

Nuclear power offers a more specific AI-infrastructure angle. The Global X Uranium ETF (NYSE:URA) held 57 securities as of Sep. 17, with Cameco Corp (NYSE:CCJ) its largest position at 22.2%.

Goolsbee’s Warning Creates New Way to Think About AI

If AI investment remains concentrated in software, chips and cloud infrastructure, traditional technology ETFs may capture most of the upside. But if the next phase of the boom increasingly runs into power, metals and infrastructure constraints, the beneficiaries could broaden.

That makes PAVE, XLU, COPX, XLE and URA potential ways for investors to examine the physical side of AI investment.

The bigger question is whether AI’s enormous capital-spending cycle ultimately becomes inflationary enough to force the Fed to keep demand constrained. Goolsbee has now explicitly put that possibility on the policy radar.

Photo: Shutterstock