Federal Reserve Governor Lisa Cook delivered a pointed warning this week. The rapid construction of artificial intelligence infrastructure now ranks among her foremost economic concerns for next year. She sees it driving price pressures that could linger far longer than many investors hope.

“The AI build out is potentially creating inflationary pressures that may not resolve very quickly,” Cook told New York Fed President John Williams during an event at the regional bank on Thursday. “So I think this is one of the main things that concerns me right now for 2027.”

Her comments, first reported by Reuters, come as inflation measured by the Fed’s preferred gauge stood at 3.4% in August. That reading has remained above the central bank’s 2% target for more than five and a half years. Cook joined a unanimous decision last month to lift the policy rate by a quarter point. The move aimed to speed the return of price stability.

Yet the governor’s analysis goes deeper. She acknowledges that artificial intelligence should lift productivity over the longer run. Many colleagues at the Fed share that view. But the timing matters enormously. “I worry about when the productivity gains that would produce disinflation will come, and where the supply bottlenecks are going to be next,” she said. Short sentences capture the tension. The gains may arrive too late.

Only days earlier Cook laid out the mechanics in prepared remarks at Oakland Tech Week. Prices for chips, computers and software have surged. Electricity and water costs each climbed around 5% over the past year, partly tied to data center demand. Core goods prices, once easing, now run above a 3% annual pace this year. Bloomberg covered those observations closely.

The investment numbers underscore the scale. Companies have announced more than $1.5 trillion in data center plans. Only a small fraction has been realized so far. That leaves substantial spending still in the pipeline. Cook noted the demand concentrates in a narrow slice of the economy. If spread evenly, the overall price index would not have climbed as much. But data centers pull on resources used across sectors. Construction labor. Energy. Skilled tradespeople. Electricians especially. An aging workforce already limits supply there.

Supply shocks compound the problem. Geopolitical tensions in the Middle East. Tariffs. Persistent effects from past disruptions. Cook observed that these shocks have proved surprisingly sticky. They shape policy thinking more than before. And AI investment adds another layer. One that shows no sign of slowing.

Her perspective aligns with other Fed voices. Governor Michael Barr recently highlighted how the surge in AI-related investment exerts a measurable effect on prices. Chicago Fed President Austan Goolsbee watches whether data center construction spills beyond its lane and overheats broader demand. These concerns have shifted the policy debate. Markets once priced in rate cuts on hopes that AI would quickly boost supply. Now the talk centers on higher rates for longer.

Productivity optimism carries risks of its own. Cook anticipates only modest disinflation from AI-driven efficiency gains within the next few years. She does not expect those benefits to arrive in time to counter broadening price pressures later this year. The labor market faces a painful transition as adoption spreads. Workers may need new skills. Firms must reorganize. Complementary investments in training and processes take time.

Some analysts push back. A Morningstar analysis argues the Fed’s response to the AI demand shock has mainly lifted interest rates while keeping output near potential. In that framing, artificial intelligence boosts rates more than inflation itself. Yet Cook and her colleagues focus on the observable data. Elevated core readings. Sector-specific price spikes. Bottlenecks reported directly by firms building the infrastructure. They describe supplies sufficient for only the next one to five years in some cases.

The governor’s caution extends to expectations. Inflation has stayed too high for too long. That raises the danger it becomes embedded in wage and price setting. Five years above target heightens that worry. So policymakers tread carefully. They weigh data meeting by meeting. Cook has stressed that addressing relative price shifts from AI is not the Fed’s role. Yet when those shifts broaden and affect the wider economy, policy must respond.

Recent coverage adds texture. CFO Brew reported on October 1 that Cook views AI as contributing to inflation now even as it promises future productivity lifts. Electricity and water increases illustrate the point. So do reports from companies that they cannot secure enough electricians or materials fast enough. The human capital constraints matter as much as the physical ones.

Investors have poured money into AI-related stocks on the belief that productivity gains lie just ahead. Equity markets reflect that optimism. But Cook’s message tempers enthusiasm. The buildout itself creates costs before it delivers benefits. Data centers must be built. Power plants constructed. Chips manufactured at scale. Those activities pull forward demand and push up prices today.

History offers limited parallels. Previous technology waves eventually raised living standards without sustained inflation. Railroads. Electricity. Computers. Each required massive upfront investment. Each eventually lowered costs. The difference now lies in speed and concentration. AI investment has accelerated faster than many forecasts. And it clusters in energy-intensive, chip-heavy segments already under strain from other shocks.

Fed officials continue to study the interplay. Speeches from earlier this year, including Cook’s own remarks in May and July, flagged the same dynamics. Announced data center spending has only grown since. The pipeline looks even larger. Uncertainty around adoption speed and productivity transmission remains high. How quickly will businesses integrate the tools? How fast will gains flow to output per worker? Those questions will shape monetary policy into 2027 and beyond.

Cook’s remarks this week crystallize a turning point in the narrative. AI no longer serves mainly as an argument for patience on rate cuts. It has become a reason for vigilance on inflation. The central bank raised rates last month despite signs of cooling in some areas. Officials cite the need for a timelier return to target. Persistent supply-side pressures from technology investment help explain why.

Markets will parse every word. Bond yields reacted to the latest comments. Equity traders weighed the implications for growth stocks. Yet the governor offered no simple prescription. She simply laid out the risks. Timing of productivity gains. Location of next bottlenecks. Potential for price pressures to broaden before they ease.

The economy has shown resilience. Growth holds. The labor market remains solid. But inflation refuses to cooperate fully. Cook and her colleagues must balance these forces without the benefit of perfect foresight. Their focus on 2027 risks signals that the AI story will influence policy long after initial hype fades. The buildout continues. So do the inflationary echoes.

And that leaves policymakers in a familiar spot. Watch the data. Adjust as needed. Hope the productivity payoff arrives before patience wears thin. For now, Cook’s assessment carries weight. The artificial intelligence investment wave brings both promise and pressure. Getting the sequence right will test the Fed in the years ahead.