Austan Goolsbee, President of the Federal Reserve Bank of Chicago, warned that inflation linked to the artificial intelligence (AI) investment boom is no longer confined to data centers and adjacent industries and is beginning to push up prices in unrelated parts of the economy. Speaking on Semafor, he said he is "increasingly nervous" that the pattern looks like plain old overheating demand.
Which costs are rising and why it matters
Goolsbee pointed to sharp price increases for land, electricity, water, construction, and the electrical components used in data centers. So far those spikes have not necessarily shown up in obvious consumer pain points such as grocery bills or haircuts, but economists are watching how wealth created by the AI boom — from paper millionaires in Silicon Valley to better-paid construction workers — is being spent.
Federal Reserve Governor Lisa Cook warned last month about "some economy-wide pressure from AI-fueled demand," though she said the boom so far looks more like a shift in demand toward AI-related sectors rather than an increase in economy-wide demand. Cook cautioned that using rate hikes to fight sector-specific inflation could be a mistake.
Limits of monetary policy against AI-driven investment
Goolsbee is not a voting member of the Federal Open Market Committee that sets interest rates; he will join the committee in 2027. Economists are trying to untangle how much of today's persistent inflation — 3.4% in August, with September’s figures due soon — is driven by higher oil prices related to the Middle East conflict, which the Fed cannot do much about, and how much comes from broader, hotter demand for goods and services.
He said that if inflation remains persistent, "it’s clear what the Fed has to do," referring to using its policy tools to curb demand. At the same time he acknowledged that those tools may be less effective against the kind of investment being driven by hyperscalers (large data-center operators). Hyperscalers and big tech firms often tolerate higher debt costs because they are engaged in a must-win strategic battle and have strong cash-generating businesses, such as advertising, that can absorb additional interest expense.
From relative price shifts to aggregate overheating
Goolsbee described how, initially, many price pressures reflected competition for inputs and resources: construction costs rose because electricians became scarce; lead times for electrical parts lengthened; and data-center buildouts bid up local factor prices. When such pressures remain confined to a subset of the economy, they represent a change in relative prices rather than aggregate overheating. But if those pressures spread into services and investment goods broadly — lifting GDP above potential and generating spillover inflation — the problem becomes aggregate overheating, which is much harder to unwind than a temporary supply shock.
What the Fed might do
Goolsbee has been critical of excessive forward guidance but also argues for demystifying the Fed’s role. He noted structural changes in transmission of monetary policy: for example, most mortgages are now 30-year fixed, reducing the speed at which an overnight rate change flows through to households. Still, the Fed’s primary tool remains the interest rate. If demand overheats, the Fed must use that tool even though it brings collateral damage.
He said his own view (reflected in his dot on the Fed’s projections) is optimistic that policy rates can eventually settle considerably below current levels if progress back toward 2% inflation is evident. He dissented against a rate cut at the end of 2025 because he did not see convincing evidence that inflationary pressures were temporary; persistent inflation since then has bolstered that stance.
Broader debates: technology, jobs and transition costs
The interview touched on wider questions of how AI will affect labor markets. Goolsbee warned against the "lump of labor" fallacy and noted historical experience with general-purpose technologies: they change jobs but typically do not drive unemployment to extreme levels, though transitions can be painful for affected workers. He suggested the speed of displacement matters greatly and said many prominent AI use cases concern higher-skilled, knowledge work rather than low-income care jobs.
He also discussed unusual labor-market dynamics such as the "low hire, low fire" environment, interpreting it as a sign of business uncertainty: firms hesitate to hire but are reluctant to fire because they fear needing to rehire once uncertainty resolves. If such patterns persist for many years, the interpretation may need to be revised.
Conclusion
Austan Goolsbee warned that AI-related demand is starting to spread beyond its initial sectors and could become aggregate overheating requiring decisive action from the Federal Reserve. He also emphasized limits to monetary policy's reach, particularly when large tech firms are less sensitive to higher borrowing costs. Policymakers remain divided over how much of current, persistent inflation is driven by AI versus supply shocks such as oil prices; upcoming data will help clarify the picture.
(Note: article synthesizes Goolsbee’s remarks and related comments from Federal Reserve Governor Lisa Cook, and reviews market and policy context as discussed in the interview.)
