The same AI investment surge the White House calls essential to America's future is now feeding inflation in electronics and threatening to drive up household electricity costs for years to come.
Prices in the federal "video and information processing equipment" category jumped 12.2% year over year through July 2026, according to a Federal Reserve Bank of Minneapolis analysis of Personal Consumption Expenditures data, the Fed's preferred inflation gauge. That same category had been falling 6.5% per year from 2015 to 2019. The reversal is stark, and the Minneapolis Fed pinned a measurable share of it on AI-driven demand for memory chips and computer hardware.
The finding matters because it puts a number on a cost that Washington has been reluctant to discuss. Goldman Sachs Research estimated in August that U.S. AI investment will total just under $600 billion in 2026. Nearly two-thirds of the S&P 500 is now tied to the AI economy, a BestBrokers report found. JP Morgan's analysis showed AI-related capital expenditures added 1.1% to GDP growth in the first half of 2025. The scale of spending is enormous, and the Minneapolis Fed's researchers concluded that "the spillovers to goods prices appear to be at least as large as tariffs in keeping core inflation high."
Core PCE, the broadest inflation measure the Federal Reserve tracks, rose 3.3% over the year in July, still well above the Fed's 2% target. The Minneapolis Fed attributed 0.4 percentage points of that 3.3% rise directly to AI-driven demand. That is a meaningful slice of the gap between where inflation sits and where policymakers want it.
The Consumer Price Index told a similar story. The Labor Department's July report showed overall prices up 0.1% from the prior month, with annual inflation cooling to 3.4%. But inside the data, the "computers, peripherals and smart home assistant devices" subcategory, identified by Dow Jones as the best proxy for AI-related prices, rose 3.5% from June to July alone.
Shikha Jain, a partner at consulting firm Simon-Kucher who leads the firm's consumer sector practice for North America, told USA TODAY that the cost pressure is cascading through the supply chain:
"Memory chip and RAM prices have increased five times because of this AI infrastructure demand, and I think that also has an impact on anything that has any kind of memory component."
Jain added that businesses facing those higher input costs are passing them along. Former Apple CEO Tim Cook told The Wall Street Journal in June that surging memory and storage chip costs had made price increases "unavoidable." When the head of the world's most valuable consumer electronics company calls higher prices unavoidable, ordinary buyers should pay attention.
Skanda Amarnath, executive director of Employ America and a former research analyst at the Federal Reserve Bank of New York, framed the dynamic in basic supply-and-demand terms:
"If one supply chain has a higher demand for a particular input, and that outstrips demand from another source, then prices go up."
That is exactly what is happening. AI companies are vacuuming up memory chips, GPUs, and related components at a pace that leaves other manufacturers competing for the same limited supply, and consumers paying more for laptops, phones, smart-home devices, and anything else that uses those parts.
The price pressure does not stop at Best Buy. Data centers, the physical backbone of the AI boom, are consuming electricity at a rate that threatens household utility bills nationwide.
The Department of Energy's Lawrence Berkeley National Laboratory found that data centers accounted for about 4.4% of total U.S. electricity consumption in 2023. The lab estimates that share could rise to between 6.7% and 12% by 2028. A 2025 study by Carnegie Mellon University and North Carolina State University projected that data centers and cryptocurrency mining together could push the average American's electricity bill up 8% by 2030. For consumers in major data center corridors like Virginia, the study estimated bills could climb 25%.
An 8% increase on an electricity bill may sound modest in isolation. Spread across tens of millions of households already squeezed by grocery and housing costs, it adds up fast, and 25% in the hardest-hit areas is not modest by any measure.
The White House has acknowledged the risk, at least on paper. President Trump announced a "Ratepayer Protection Pledge" earlier in 2026, asking AI and tech companies to "protect American consumers from price hikes" caused by the buildout and to "lower electricity costs" for households over time. The pledge also asks companies to build or buy new power supplies, invest in local jobs, and make surplus energy available to communities to prevent blackouts.
Google, Microsoft, Meta, Oracle, xAI, OpenAI, and Amazon all signed. Governors from 23 states joined in July. Trump praised the effort in August, telling reporters:
"I'm very proud of it. It was my idea. They're gonna build their own electric producing facility and that electricity that they have left over is going to the grid."
The problem is that the pledge is non-binding. No company faces penalties for breaking it. No regulator is tasked with measuring compliance. Amarnath was blunt about its limits:
"It's not as if the signatories to these things are actually being held liable for anything."
He went further, saying that anyone serious about protecting ratepayers "would have to develop some frameworks for saying this is the accounting for this stuff, this is how we decide that this data center is imposing this cost and needs to pay for it." Without that framework, the pledge amounts to a handshake between the most powerful companies on earth and the government that is subsidizing their growth, with taxpayers left to hope the handshake holds.
Federal Reserve Chair Kevin Warsh, who took over earlier in 2026, has positioned himself as an optimist on AI's long-term economic potential. In a November op-ed in The Wall Street Journal, he called AI "a significant disinflationary force" and argued that "a 1-percentage-point increase in annual productivity growth would double standards of living within a single generation."
But at the Jackson Hole economic symposium on August 28, Warsh's tone was more cautious. He posed two questions without answering them:
"Will the application of AI cause a significant, sustained rise in productivity across the economy? And if so, when?"
"Will token usage be complimentary or competitive to labor?"
Warsh has created five task forces at the Fed focused on productivity and jobs, with recommendations expected later this year. The questions he raised at Jackson Hole are the right ones. But for now, the disinflationary payoff he promised in his op-ed remains theoretical, while the inflationary costs are already showing up in government data.
The White House's AI Action Plan frames the investment push in sweeping terms, declaring that "whoever has the largest AI ecosystem will set the global standards and reap broad economic and security benefits" and promising "a new Golden Age of innovation, human flourishing, and technological achievement for the American people."
The ambition is understandable. Ceding AI leadership to China would be a serious strategic mistake. But ambition does not exempt policymakers from accounting for the costs that land on working families right now.
A Reuters/Ipsos poll found that 53% of American adults fear AI could cost them or someone in their household a job. About half of U.S. adults already use AI chatbots, according to Pew Research. The technology is spreading fast, and anxiety about its effects is spreading with it.
Amarnath challenged the assumption that productivity gains from AI will automatically lift wages:
"Is it the case that more productive societies can pay their people better? Yes. Is it the case that we see when productivity rises, real wages also rise? No....Even if you assume that productivity is going to increase, it's not a given that's going to show up as better conditions for workers."
That gap, between what AI could do for the economy in theory and what it is doing to household budgets in practice, is the central tension the administration has not resolved. ConnectOne Bank founder and CEO Frank Sorrentino has explained Warsh's logic that AI could eventually bring prices down by making businesses more efficient. Amarnath offered a concrete example: AI could force tax preparation services to slash fees or find new ways to deliver value. In that scenario, consumers win.
But "eventually" is doing a lot of work. Memory chip prices have already increased fivefold. Electronics costs are rising at double-digit rates after years of decline. Electricity bills in data center corridors are projected to jump by a quarter. And the only formal protection the White House has offered consumers is a voluntary pledge with no enforcement mechanism.
The administration deserves credit for recognizing that AI dominance matters for national security and economic competitiveness. The scale of private investment flowing into the sector is a sign of confidence in the American economy. And if Warsh is right that AI eventually becomes a disinflationary force, the short-term pain could yield generational gains.
But none of that changes the arithmetic facing a family in Virginia whose electricity bill may jump 25% by the end of the decade, or a parent buying a laptop for a college-bound kid and finding prices 12% higher than last year. The Ratepayer Protection Pledge asks trillion-dollar companies to voluntarily absorb costs they have every financial incentive to pass along. History suggests how that usually ends.
The AI buildout may well be worth the investment. But telling Americans to trust a handshake while their bills climb is not a serious answer, and the people paying those bills know it.