AI Buildout Adds to Inflation, Splits the Fed
Get the Finance newsletter
Daily finance — markets, central banks, M&A, the prints that move money. Free.
- U.S. AI capital expenditure is projected to hit $581 billion in 2025 and as much as $1 trillion globally, rising to 2.8% of GDP by 2028, per Goldman Sachs Research.
- Federal Reserve officials voted in July to hold rates at 3.5%-3.75%, with Minneapolis Fed President Neel Kashkari dissenting in favor of a hike, explicitly citing data-center-driven price pressures.
- Household electricity prices rose 10.1% over the two years through June 2025, outpacing the broader 6.3% CPI increase, as power-hungry AI data centers strain grids.
- DRAM prices are projected to surge 400% by year-end versus 2024, per JPMorgan Chase estimates, while computer software and accessories costs are up 22.9% since June 2024.
- OpenAI chief economist Ronnie Chatterji said power users deploy AI at eight times the rate of average companies — up from a 2x gap just three months ago — highlighting concentrated rather than broad adoption.
- Fed Chairman Kevin Warsh appointed Stanford's Charles Jones (currently on leave at Anthropic) and venture capitalist Marc Andreessen to a task force advising the central bank on AI's economic effects.
- Peter Boockvar of One Point BFG noted the internet era produced only a 1.5% productivity gain over 30 years, questioning whether generative AI will outperform that historical baseline.
Why it matters: Households are already paying for the AI buildout — electricity up 10.1%, DRAM up 400% — while the Fed, which held rates at 3.5%-3.75% in July, is now internally split over whether AI-driven costs warrant tighter policy. Kashkari's explicit dissent for a rate hike, paired with Warsh's pivot from AI-as-deflation advocate to cautious observer, means monetary policy now directly trades off against the tech sector's multi-trillion-dollar bet.
Ask SkimNews
