Fed’s Barr says AI’s economic payoff hinges on competition versus concentration

Federal Reserve Governor Michael Barr has a simple framing for one of the most complicated questions in economics. He says the effects of artificial intelligence depend partly on whether the technology increases market concentration or competition.

Two roads from the same technology

Barr’s point is that AI is not a single outcome waiting to happen. It is a fork in the road.

Down one path, the gains pool among a handful of dominant firms. Barr has used the term “hyperscalers” for these players. Picture the massive tech companies with the cloud capacity and capital to train and run AI at enormous scale.

Down the other path, the benefits spread more widely across the economy. Smaller companies adopt the tools, compete harder, and pass gains along to customers and workers.

Which road the economy takes is not purely up to the technology. According to Barr, regulatory frameworks and competition policy play a large role in deciding whether benefits concentrate or disperse.

The adoption gap is already wide

Barr has pointed to data that suggests the concentration question is not hypothetical. As of 2024, about 75% of large companies were using generative AI.

Smaller businesses, by contrast, showed adoption rates in the high single digits.

Barr has framed the range of possible effects as a spectrum. At the mild end, AI simply augments individual tasks, helping workers do their jobs a bit faster. At the extreme end, it disrupts entire industries.

He has discussed scenarios running from incremental to revolutionary. The speed of adoption and the size of productivity gains help determine where on that spectrum the economy lands.

The $3 trillion construction project

AI is not just software. It runs on physical infrastructure, and lots of it.

Barr has pointed to an expected surge in data center investment estimated at around $3 trillion. That spending could lift productivity over time.

It could also create localized inflationary pressure in the near term. Barr flagged sectors like electricity and construction as places where prices may feel the strain.

Where markets come in

Barr’s concerns extend beyond the real economy into financial markets themselves. In a November 11, 2025 address at the Singapore Fintech Festival, he described AI’s potential to transform economies while warning about the risks.

Among them: AI-powered trading algorithms that could lead to tacit collusion and heightened market volatility.

Barr has also flagged market herding as a risk. If many firms rely on similar models trained on similar data, they may all buy or sell at the same moment.

His list of concerns also includes bias in lending. AI models used to approve loans could bake in patterns that disadvantage certain borrowers, even if no one designed them to.

Through all of this, Barr has stressed the need for balance. Regulation should manage these risks without smothering the innovation that makes AI valuable in the first place.