Business

AI data center boom lifts US electricity prices 10.1% in two years

Published Aug 12, 2026, 11:12 AM2 min readNewUJ Editorial Desk

AI data center boom lifts US electricity prices 10.1% in two years
Photo: Parrish Freeman · Unsplash
0 0
XWhatsAppTelegramLinkedIn

The massive buildout of artificial intelligence infrastructure is driving up costs and complicating the Federal Reserve's fight against inflation, according to economists and company executives. While tech leaders have promised deflationary benefits, the immediate reality is higher prices and slow adoption.

Companies across the U.S. are spending heavily on data centers and AI hardware, with capital expenditure expected to reach $581 billion this year, or 1.8% of GDP, according to Goldman Sachs Research. That share is projected to rise to 2.8% by 2028. The spending spree is straining supply chains and pushing up electricity costs, with household electricity prices rising 10.1% in the two years through June, outpacing the overall 6.3% inflation rate.

The inflationary pressures are creating a dilemma for the Fed. In July, officials voted to leave interest rates unchanged at 3.5% to 3.75%, but the decision was not unanimous. Minneapolis Fed President Neel Kashkari dissented, citing the need to restrain AI-driven price increases. Fed Chairman Kevin Warsh, who previously argued AI would be disinflationary, has adopted a more cautious tone, saying the timing and magnitude of supply-side effects remain hard to predict.

Adoption of AI by businesses remains limited, with a Census Bureau survey in May showing only 17% to 20% of firms using it, and large companies far more likely to do so. OpenAI chief economist Ronnie Chatterji said power users deploy AI at eight times the rate of average companies, a gap that has grown from two times three months earlier. But he acknowledged it will be a while before productivity gains show up in statistics.

Some executives warn that integrating AI into large organizations is challenging. Julie Averill, former chief information officer at Lululemon, said getting people to change behaviors and trust models is hard, even when the technology works. Stanford professor Charles Jones, now on leave at Anthropic, points to "weak links"—tasks that cannot be easily automated—as a barrier to broader productivity gains.

Warsh has appointed Jones and venture capitalist Marc Andreessen to a task force that will inform the Fed's thinking on AI. Their report, due in a few months, will feed into a debate over whether AI will ultimately curb or fuel inflation. For now, the costs are tangible: computer software and accessories prices have risen 22.9% since June 2024, and DRAM costs are estimated by JPMorgan Chase to jump 400% by year-end compared to 2024.

Sources

Report / request removal

Related

Comments

No comments yet. Be the first.