Richmond Fed's Barkin: The 'AI Apocalypse' Hasn't Arrived

The Federal Reserve raised interest rates last week for the first time since mid-2023, and Richmond Fed President and CEO Tom Barkin says the forces behind that decision — tariffs, gasoline prices and an AI investment boom inflating technology equipment costs — are proving harder to shake than policymakers hoped.

Speaking Thursday at an Economic Club of Washington, D.C., event in conversation with Barbara Humpton, CEO of USA Rare Earth, Barkin said the inflation picture has changed over the past six months. Earlier this year, he argued, it was easier to claim the persistence was temporary, that oil prices, tariffs and the AI buildout would eventually "morph" away.

"If inflation's not going to come down relatively quickly," Barkin said, "then you have to look in the mirror and say inflation looks like it's been here for a while."

A log flume for the new Fed chair

Barkin offered a vivid metaphor for the job facing the incoming Fed leadership. He described it as a log flume ride: "You go slowly up the hill of the last ramp, and then you go full speed down into a pool of water, and the pool of water comes and soaks the guy in front. That's Kevin Warsh or Jay Powell. The rest of us are in back. Our hands are in the air."

The image captures the bind: the Fed has moved back toward higher rates as inflation has proved more persistent than expected, while an enormous wave of AI investment is reshaping prices and the labor market simultaneously.

Barkin, who served as CFO of McKinsey before joining the Richmond Fed, said he has "the deepest respect for anyone who would be foolhardy enough to take on the leadership of the Federal Reserve System," and added that Warsh "made a number of, I think, very profound calls."

An enormous investment cycle

On AI, Barkin described an economy caught in a massive investment cycle. Companies are spending aggressively because they expect AI to lift productivity — and that same expectation is making them more cautious about hiring.

The technology is already at work in coding, call centers, compliance paperwork and engineering, Barkin said. Yet strong corporate earnings are working against layoffs. Companies sitting on deep backlogs of technology projects would rather redeploy freed-up capacity than cut headcount, according to Barkin.

The harder problem, he argued, is organizational. "The unit of an AI-enabled task is not precisely the same as a mid-level manager and what they do," he said. Companies can see how AI makes individual tasks more efficient without knowing what that means for the people performing them. Many firms have not yet figured out how to translate task-level efficiency into a redesigned staffing model.

Barkin also raised a less obvious possibility: AI could make it easier to bring inexperienced workers into skilled trades. He pointed to auto mechanics, where an AI assistant could hand a less-experienced worker a real-time checklist to compensate for missing expertise.

Not the apocalypse

"I think there's a lot of talk about the AI apocalypse," Barkin said. "We're clearly not yet there." He added that he is hearing "a lot less fervency in the conviction that we're going to be there" than there was six months ago.

Humpton, whose company is focused on rebuilding a domestic rare-earth workforce, offered a real-world example of the other side of the transition: retraining workers for jobs that increasingly demand new combinations of technical and practical skills. She called the effort "truly transformative."

The takeaway for executives: with the Fed back in tightening mode and AI spending itself feeding into equipment prices, the technology wave Barkin describes is now both a productivity bet and an inflation variable the central bank cannot ignore.