Article on # Furious pace of AI investmen...

Federal Reserve officials are starting to ask whether the surge of investment in artificial intelligence is creating risks for the financial system. The question is moving from a fringe concern to formal discussion as data-center spending, debt-financed buildouts, and uncertain returns draw the attention of central bankers. New York Fed President John Williams said in an interview last Friday that he doesn't see a bubble forming. "What we're seeing is a very high level of excitement, enthusiasm around new technology, around AI," Williams said. He acknowledged that investors are "trying in real time to solve an almost intractable problem": how large AI's benefits will prove to be. Answering that question will cause volatility, he said. Williams said the increased borrowing supporting AI investment is being managed by companies with strong earnings. "I'm not as worried about the financial stability from the leverage right now," he said.

Some Fed officials see reasons for caution

Not everyone at the Fed shares that view. Kansas City Fed President Jeff Schmid said in a Tuesday speech that the industry's financing model deserves closer scrutiny. "I would argue that there's some signs that we have to really start to talk about that on a macro level, is this industry becoming another too big to fail?" Schmid asked. He pointed to the chain of commitments linking data centers, energy providers, and the communities they serve, and asked whether that circle is becoming over-leveraged. "If you get a spark that starts a flame, what happens?" Schmid said. San Francisco Fed President Mary Daly said Wednesday that the raw numbers are hard to ignore. Looking at the growth rate and amount of AI investment, you could easily call it "very worrisome," she said. But she also noted that many commitments are still announcements, not physical assets, which reduces the risk of stranded investments if the sector contracts. She flagged the rise of borrowing to fund growth as a potential issue.

Scale of the buildout

The data-center buildout remains smaller than the housing boom that preceded the 2008 crisis. Torsten Slok, chief economist at Apollo, said the buildout is less than half the size of the housing boom, which peaked at 6.6% of GDP in 2005. The current investment pace relative to GDP is growing faster than housing did before the crisis.

Why this matters for finance professionals

The Fed's attention signals that AI investment risk is now a macro factor, not just a tech-sector story. The central bank is building what Daly described as "a dashboard" of potential failure points, not the problems of the last crisis but what could tip the current one. For finance professionals, the practical question is how these discussions shape credit conditions and market expectations. The Fed's approach, tracking what could go wrong rather than what already has, is a useful framework for AI for Finance practitioners assessing risk in their own portfolios. Executives weighing AI investments face the same question the Fed is asking: how much risk is acceptable in pursuit of a new technology. That question is central to AI for Executives & Strategy.