Fed’s Lisa Cook warns AI spending could fuel inflation into 2027

Artificial intelligence was supposed to make everything cheaper. According to Federal Reserve Governor Lisa Cook, it may make some things more expensive first.

Cook has flagged AI-driven capital spending as a source of inflationary pressure. She says that pressure could become a primary concern for the central bank heading into 2027.

What Cook actually said

Cook laid out her case in a speech at Oakland Tech Week on September 28, 2026. Her focus was the gap between what companies have promised to spend on AI and what they have spent so far.

Companies have announced approximately $2 trillion in AI-related investment plans, much of it aimed at data centers. Only a minor portion of that money has actually gone out the door.

Cook pointed to costs tied to the buildout, including electricity. She noted that key inputs have each risen around 5% over the past year.

Cook did not frame AI as purely inflationary, though. She expects productivity gains from the technology could produce modest disinflation “within the next few years.”

The catch is timing. She cautioned that those gains are unlikely to offset the inflationary pressures she expects to persist through late 2026.

The 2027 warning

Cook returned to the theme on October 1, 2026, in a discussion with New York Fed President John Williams. There, she pointed to 2027 as the year AI’s inflationary effects could take center stage.

Her reasoning rested on two factors. The first is ongoing supply bottlenecks in the resources AI infrastructure depends on. The second is uncertainty over when productivity relief will actually show up.

Cook also drew a firm line on what the Fed should do about it. She said monetary policy should not try to target inflation in specific sectors. Her explanation was simple: the Fed’s tools are too blunt for that kind of precision work.

What this means

The most direct implication is for the industries supplying the AI buildout. Utilities, construction and technology hardware all sit in the path of the cost pressures Cook described.

The roughly $2 trillion in announced but largely unspent commitments is the number to watch. If that money flows out on schedule, it will keep pressing on the same constrained supply chains Cook highlighted.

Because Cook ruled out using rate policy to target sector-specific inflation, the Fed is unlikely to tighten specifically to cool AI spending. The more relevant question is whether AI-related cost increases bleed into the broader inflation picture.

Track any early evidence that AI is delivering the productivity gains she expects. That is the variable that could change the timeline she laid out.