Fed Vice Chair Jefferson Warns AI Investment Could Raise Inflation Before Productivity Benefits Emerge
Federal Reserve Vice Chair Philip Jefferson said the surge in AI-related investment is currently adding to demand and inflation, while any broader productivity relief is likely to arrive later and remains uncertain.

Federal Reserve Vice Chair Philip Jefferson said Thursday that the rapid buildout around artificial intelligence is creating a difficult near-term inflation problem for policymakers because the investment boom is lifting demand now, while the technology’s broader productivity benefits may take longer to arrive.
Speaking at an International Monetary Fund and Central Bank of Brazil conference, Jefferson described AI as a shock that is affecting both sides of the economy at different speeds. Optimism around the technology is fueling spending on data centers, advanced computing equipment and related products, he said, even as the eventual supply-side gains from automation and higher productivity remain prospective rather than fully visible in the data.
That sequencing matters for the Federal Reserve. A technology wave that raises productive capacity can eventually help contain inflation. In the interim, though, heavy investment and stronger demand for scarce inputs can add to price pressure, especially when inflation is already above target and supply in key areas is struggling to keep up.
AI buildout is adding to demand and price pressure now
In his Nov. 19 remarks, Jefferson said business adoption of AI has risen considerably over the past two years. On the demand side, he said, enthusiasm for the technology has been boosting business investment in data centers, advanced computing equipment and related products. On the supply side, AI is likely to automate some tasks and augment others, which could eventually raise productivity and the growth rate of potential output.
His central warning was about the gap between those two effects. Jefferson said the AI buildout is currently exerting upward pressure on demand and inflation. That leaves policymakers facing a familiar but awkward problem: a shock that can support long-run growth, but that may worsen price pressure before the economy sees the full productivity payoff.
He also tied the issue to broader resource constraints. When large sums are funneled quickly into a narrow set of inputs, prices can rise sharply in those areas before new supply comes online. Jefferson highlighted business spending on the infrastructure needed for model training and inference, and he noted that sectors such as advanced chips have had difficulty keeping pace with demand. For the Fed, that means the AI boom is not just a story about innovation. It is also a story about near-term resource utilization and the general price level.
The speech did not announce a policy move, and Jefferson did not attach a timetable to when productivity gains might outweigh those inflationary effects. Instead, he framed AI as one of several overlapping shocks that the Fed must assess alongside energy prices and trade-policy changes. The message was less about making a directional call on the next meeting and more about showing why the inflation outlook remains complicated.
Productivity gains may come later, but Jefferson said the evidence is still limited
Jefferson’s warning on Thursday builds on a longer speech he delivered on Nov. 13 about technology, productivity and the macroeconomy. In that earlier discussion, he argued that AI has the potential to lift the economy’s productive capacity substantially, but gradually. He compared the current moment with earlier waves of information-technology investment, when productivity benefits took time to show up in aggregate statistics because firms first had to experiment, reorganize workflows and broaden adoption.
That distinction between capability and economy-wide payoff runs through his recent comments. Jefferson said recent surveys suggest AI use is spreading, but not uniformly. In the Nov. 13 speech, he pointed to Census Bureau survey evidence showing that as of October, roughly 25% of U.S. firms were using AI in at least one business function around midyear. Adoption was materially higher in knowledge-intensive industries than in sectors such as mining or leisure and hospitality.
He also noted that surveys of workers show 40% to 50% of working adults using generative AI on the job, but only a modest share of those users rely on it every day. That is one reason he has been cautious about claiming that the recent pickup in productivity growth can already be pinned on AI. Jefferson said labor productivity growth averaged 2.2% between 2020 and 2025, up from a 1.5% average over the prior business cycle, but he also said that improvement began before popular AI chatbots emerged in 2022 and likely reflects several forces, including stronger business formation.
In other words, the broad productivity benefit that could ultimately make AI disinflationary is still more of an informed expectation than a measurable present-day fact. Jefferson said the evidence so far appears consistent with only a limited contribution to aggregate productivity, even though the technology has the potential to contribute much more over time. The magnitude and timing of those gains, he said, remain highly uncertain.
Labor-market disruption and policy calibration remain part of the Fed’s challenge
Jefferson also emphasized that AI’s economic effects will not be confined to output and prices. He said the technology could displace some workers even as it creates new jobs and new firms. That transition could be painful for households whose roles are exposed to automation, and it could leave lagging businesses at a competitive disadvantage if they fail to adapt.
For monetary policy, that raises a second layer of complexity. If AI eventually boosts productivity on a sustained basis, it could also lift firms’ appetite for investment and increase demand for funding, which in turn could push up the economy’s neutral rate of interest. But Jefferson cautioned that estimating such shifts in real time is difficult. The relationship between productivity growth and long-run real rates has historically been noisy, and policymakers cannot assume that every burst of technological enthusiasm will translate neatly into higher sustainable growth.
His broader policy framework remains data dependent. In the Nov. 13 speech, Jefferson said inflation was still notably above the Federal Open Market Committee’s 2% target, with the personal consumption expenditures price index up 3.6% in September and core PCE inflation at 3.1%. He also said he had joined the majority of colleagues at the prior FOMC meeting in voting to keep the federal funds rate steady. Since then, he said, inflation readings had continued to show persistent price pressure, making the inflation trajectory the more important signal for judging the appropriate stance of policy.
That backdrop helps explain why the AI story matters to the Fed now. If the technology eventually boosts supply enough to cool inflation, that would be welcome. Until those gains are broad and measurable, however, the spending boom around AI infrastructure can work in the opposite direction by raising demand and costs. Jefferson’s message was that central bankers have to live in that uncomfortable overlap: a period in which a potentially productivity-enhancing technology is helping support growth, but is also complicating the path back to price stability.
Latest News
View all news- Walmart Raises Full-Year Outlook After Revenue Climbs 5.2% in Third Quarter
- RBI Absorbs ₹11,682 Crore Through Reverse Repo Auction at 5.49%
- U.S. Leading Economic Index Falls 0.3% in October as Housing Permits Weaken
- U.S. Jobless Claims Rise to 203,000 as Continuing Claims Jump by 70,000
- Williams-Sonoma Reports 21% EPS Growth, Helped by Tariff Refunds