Powell: AI Buildout 'Probably Pushing Inflation Up' as Fed Holds Rates
In a significant statement on March 22, 2026, Federal Reserve Chairman Jerome Powell identified the explosive growth of artificial intelligence infrastructure as a new source of upward price pressure. Speaking after the Fed voted to hold its benchmark lending rate steady, Powell argued that the immediate economic impact of the AI boom is inflationary, challenging the prevailing narrative that AI's productivity gains would quickly lower costs.
In the short term, what’s happening is we’re building data centers everywhere, and that’s actually putting pressure on all kinds of goods and services that go into building these things. So that’s actually probably pushing inflation up.
— Jerome Powell, Federal Reserve Chairman
Powell's logic separates the AI revolution into two phases. The first, which is occurring now, involves a massive physical buildout that increases demand for construction materials, labor, and, most critically, energy. The second, more distant phase involves the disinflationary productivity benefits from AI itself. This perspective suggests that near-term interest rates may need to remain higher for longer to counteract this new demand-side pressure, even as the Fed revised its long-run growth estimates up from 1.8% to 2.0%.
Grid Strain Fuels $31B in Utility Rate Hikes
The Chairman's concerns are substantiated by mounting stress on the U.S. power grid and supply chains. Utilities requested a record $31 billion in rate increases in 2025—more than double the prior year—to fund grid upgrades necessary to support power-hungry data centers. This trend is already expected to hit consumers, with Goldman Sachs warning that electricity prices could rise by 6% between 2026 and 2027.
This demand is so intense that it has created its own bottleneck. A report from Wood Mackenzie found that data center development is now slowing, not from a lack of demand, but because the electrical grid cannot keep pace. According to the report, only one-third of data center projects currently in the pipeline are under active development, highlighting the physical constraints on the AI expansion.
Investors Pivot to AI Bottleneck Stocks
Powell's analysis provides a clear playbook for investors looking to navigate this new dynamic. The strategy shifts from broad AI software bets to companies that control the physical bottlenecks of the infrastructure buildout. These firms possess significant pricing power in an environment where demand far outstrips supply.
Key areas of focus include providers of critical raw materials and components. Copper producers like Freeport-McMoRan are essential for wiring the new data centers. In the semiconductor space, Micron Technology, one of only three major suppliers of high-bandwidth memory (HBM) chips, has already sold out its entire 2026 supply. The most acute bottleneck, however, is energy. Power producers, particularly nuclear operators like Constellation Energy, are uniquely positioned to benefit from the relentless demand for electricity, marking them as key beneficiaries of this inflationary phase of the AI boom.