The US economy has a new load-bearing wall, and it’s made of GPUs. John C. Williams, President of the Federal Reserve Bank of New York, said that without the massive wave of artificial intelligence investment currently flowing through the economy, overall demand and output would be significantly weaker.

How much weaker? Williams suggested that growth could have been reduced to roughly one-third of its actual pace during certain periods. Strip out the data centers, the AI infrastructure buildouts, and the capital expenditures from Big Tech, and the American economy starts looking a lot less impressive.

Hundreds of billions keeping the lights on

The numbers behind the AI investment boom are staggering. Williams pointed to estimates placing AI-related capital expenditures in the hundreds of billions of dollars annually. That spending is coming primarily from major technology companies racing to build out the data center capacity and infrastructure needed to train and deploy increasingly powerful AI models.

The Fed’s AI dilemma

Williams’s comments sit within a broader, more complex debate happening inside the Federal Reserve. The central bank is trying to figure out what AI means for two things it cares deeply about: inflation and the neutral rate of interest.

Williams’s own models peg the neutral rate, known as r*, at around 1.1%. That’s the theoretical interest rate where the economy is neither being stimulated nor restrained.

In the near term, hundreds of billions in AI spending acts as a demand stimulus. Companies are pouring money into infrastructure, which pushes up demand for labor, materials, and energy. But the long-term story could be the opposite. If AI delivers on even a fraction of its productivity promises, it could be profoundly disinflationary. More output per worker means lower unit costs.

The Fed is debating whether AI investment will yield net inflationary or disinflationary pressures, with ongoing discussions projected to continue into 2025 and 2026 as officials evaluate near-term demand boosts against long-term productivity effects.

What this means for investors

AI is no longer just a growth narrative for tech stocks. It has become a macro variable. When a senior Fed official says the economy would be meaningfully weaker without AI spending, that changes how you think about the durability of current growth rates.

Watch the capex numbers. Quarterly earnings reports from major tech companies now carry macroeconomic significance. When Microsoft, Google, Amazon, or Meta report their capital expenditure plans, they’re giving you a preview of a meaningful chunk of US economic demand. Any sign of pullback in those numbers should be treated as a macro signal, not just a sector signal.