"Surging electricity bills" become a focal point in the US midterm elections, as AI data centers stand at the "political volcano crater."

Soaring electricity costs are evolving into a core issue on the U.S. political agenda, with increases outpacing other types of inflation and pushing utility bills to the forefront of political discourse ahead of the midterm elections. Data centers, due to their huge power consumption, have become targets, and politicians from both parties are wading into the debate over rising electricity prices, leveraging the issue as a pain point for voters.

According to the Wall Street Journal on Saturday, the Trump administration last Friday convened officials including the governors of Pennsylvania, Ohio, and Virginia at the White House to push the nation's largest grid operator to hold an emergency electricity auction. The administration demanded that large tech companies provide their own power supply or bear the costs of building new power plants to ease concerns that their data centers are driving up electricity prices for ordinary Americans. Trump praised Microsoft's pledge to pay higher electricity prices for all its U.S. data centers on social media, "ensuring Americans don't 'foot the bill' for their power consumption."

In December last year, U.S. electricity costs rose 6.7% year-over-year and are up about 38% cumulatively since 2020, while overall consumer prices have increased just 2.7% over the same period. This gap has made electricity prices a key issue in gubernatorial races across 36 states this year and could affect the outcome of utility commission elections in 9 states.

Faced with this political risk, Goldman Sachs advised investors to hedge against "AI politicization" risk. The bank's trading desk pointed out that as the midterm elections approach, policymakers' public anxiety about data center energy consumption is rising, and AI infrastructure could become the next target of unexpected government policy shocks.

Electricity price hikes far outpace overall inflation, politicians seize voter pain points

The reasons for rising electricity costs are varied and complex. According to research led by Lawrence Berkeley National Laboratory, factors such as aging equipment replacement, rebuilding after storms and fires, and state renewable energy programs have all contributed to price increases. Fluctuations in fuel costs, like natural gas, can also push bills up or down.

Political pressure is spreading across the states. Indiana's Republican governor Mike Braun said last September, "We can’t stand it anymore," and instructed the state consumer advocate to evaluate utility company profits. Maine's Democratic governor Janet Mills called an electricity rate hike request "out of touch with reality" and encouraged the state utility regulator to reject the proposal, which the commission did last November.

Last October, Republican Senator Josh Hawley wrote to the CEO of a utility company, saying, "Families across Missouri are struggling to pay their bills." His concerns included rising energy prices and ensuring that residential customers don’t have to pay for the massive electricity use of new data centers.

Energy provider Ameren Missouri responded by saying its residential electricity prices are among the lowest nationwide but acknowledged Hawley's concern: "Rising household costs are a real challenge." A new Missouri law seeks to ensure large customers bear the grid costs they create.

Data centers become scapegoat for rising electricity prices

Data centers are bearing much of the blame for surging electricity costs. The core dispute is over how costs are distributed. Charles Hua, executive director of the nonprofit PowerLines, raised a key question: "Which costs should be borne by residential consumers like you and me, and which should be borne by large commercial clients like tech companies or big industrial customers like manufacturing plants?"

Trump has folded these concerns into a series of recent proposals on living costs. He praised Microsoft’s commitment to pay higher electricity rates for all its U.S. data centers powering AI models, saying it "ensures Americans don’t 'foot the bill' for their power consumption."

According to the Wall Street Journal, parts of the Northeast and Mid-Atlantic have seen cumulative bill inflation of 29% over the past three years (20 percentage points higher than CPI). The causes are multifaceted: electrification, data centers, and manufacturing returning home have pushed demand higher; coal plant retirements have tightened regional markets, raising capacity prices passed on to customers; volatile natural gas prices and domestic supply constraints in some areas; increased public welfare charges to support state and federal energy efficiency, wildfire mitigation, and clean energy goals; and higher delivery charges due to utility company investment in aging grid infrastructure. In lower-inflation states, abundant local resources (coal, wind, natural gas) help maintain low prices and ample supply.

Midterm elections push electricity prices to the political forefront

For utility companies, 2026 seems to be a tough time to seek rate hikes. John Bartlett, president of industry investor Reaves Asset Management, said, he expects electricity costs to become a focus in gubernatorial races in 36 states this year, though he noted governors aren’t responsible for all electricity price drivers, including one critical input: natural gas costs.

Most state utility commissions are appointed, but in the 10 that are elected, 9 have an election this year. Hua anticipates fierce contests in Arizona, Louisiana, and Georgia. Affordability and data centers are issues in all three, with Georgia voters ousting two incumbent commissioners last November.

Ed Hirs, an energy economist at the University of Houston, said: "The general rule is, if gasoline at the pump or electricity at the meter goes up, incumbents don’t get reelected. That applies to any incumbent, whether Democrat or Republican." However, he noted that many factors driving up electricity prices have been brewing for years, especially in competitive markets, where he believes generators lack sufficient financial incentive to invest enough capital in plant maintenance or new builds, but now additional generation is needed.

Goldman Sachs recommends hedging "AI politicization" risk

Goldman’s trading desk has been responding to client questions about trades spreading from AI to other parts of the stock market. The firm identified three main concerns: companies have invested massive cash flow into data center infrastructure, raising persistent worries about expected investment returns; uncertainty over whether data center capacity needs are accurately measured and if there’s oversupply; and midterm elections may introduce regulatory controls on data center siting and utility bill inflation.

Goldman’s trading desk does not expect the first two issues to drive an AI sector selloff in 2026. Its report stated that profit and valuation analyses for all AI categories show pricing is reasonable and not comparable to historic bubbles (Internet bubble/global financial crisis). With the exception of Google, the Mag5 group has underperformed, but earnings growth remains healthy.

But Goldman does expect this topic to be in the spotlight for the midterms. Several U.S. policymakers have publicly expressed concerns over data center energy consumption. Massachusetts Senator Edward J. Markey wrote to the Federal Energy Regulatory Commission (FERC) last November, urging it to prevent data center growth from "sharply driving up energy costs for American households," pointing out that data centers are fueling electricity demand and could burden families by hiking utility bills. The letter was co-signed by six other senators. Florida Governor DeSantis backed a change to state energy law in 2024 to emphasize cost and reliability goals, and throughout the summer speeches targeted energy use driven by data centers and AI.

Goldman's equity research team explained that accelerating utility bill inflation in some areas has triggered concerns over customer affordability. The Trump administration’s policies have already had unexpected effects on the stock market—will AI be next? The best-performing themes in equities have consistently been led by AI infrastructure leaders.

Goldman proposes three preferred trading strategies: go long on non-tech and non-AI firms leveraging AI to boost productivity (AI Productivity Basket GSXUPROD); go long on Mag5 excluding Google (GSCBM5XG); go long on AI volatility (GSVIAIV1) to hedge "AI politicization" risk as midterms approach. The firm’s favorite hedge is going long on AI volatility.

Risk Warning and DisclaimerThe market involves risks; investments should be made prudently. This article does not constitute individual investment advice nor take into account the special investment objectives, financial situations or needs of any particular user. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their specific circumstances. Investing accordingly is at your own risk.