• Federal Reserve Governor Christopher Waller said there is nothing "bubble-like" in how AI is being priced.
  • His remarks contrast with recent warnings from U.S. and European regulators about the financing behind the AI boom.
  • The debate over AI valuations and debt-fueled infrastructure spending continues to simmer.

Federal Reserve Governor Christopher Waller on Wednesday pushed back against concerns that the rapid ascent of artificial intelligence stocks resembles a speculative bubble. "There is nothing 'bubble-like' in how AI is being priced," Waller said, according to a statement. The comment offers a counterpoint to a growing chorus of regulators who have flagged potential risks in the financing of the AI infrastructure buildout.

The debate over AI valuations has intensified this month. On November 13, the Federal Reserve's Financial Stability Report noted elevated stock valuations and a surge in AI-related borrowing, with more than $680 billion in advanced-computing debt issued in the first ten months of 2026—roughly five times the amount issued in all of 2023. AI was the most frequently cited risk among the Fed's surveyed market contacts, with 87% citing it, up from 50% in the spring.

Just days later, Fed Governor Michael Barr said balance-sheet debt at seven of the largest AI firms was expected to exceed $800 billion by year-end. The Fed also estimated that the potential maximum nominal value of commitments such as leases, long-term contracts, and guarantees exceeded $3 trillion, though that figure is not recorded debt or an estimate of likely losses. On Wednesday, the European Central Bank warned that an AI selloff could spread from stocks to bonds and reach euro-area investors, as more AI-related borrowing is issued in euros.

Waller's remarks suggest that equity valuations may be justified by expected profits, even as the borrowing behind data centers could amplify losses if those profits disappoint. The distinction is important: regulators have not declared that AI stocks are in a bubble, but they are increasingly focused on the financial plumbing that supports the buildout.

The exposure extends beyond tech shareholders. Data-center developers depend on financing and customer contracts, while lenders and bondholders face credit risk. A severe reversal could ripple through the broader economy, affecting investment, spending, and employment.

The ECB drew parallels to the dot-com era and the subprime mortgage crisis, though it cautioned that neither comparison is a prediction. The dot-com example shows how a transformative technology can still attract too much infrastructure investment, while the subprime comparison highlights how opaque obligations and supplier financing can obscure who ultimately bears the losses.

In the near term, market participants are watching whether AI revenue keeps pace with infrastructure spending and whether lenders demand more compensation for financing it. Fed Vice Chair for Supervision Michelle Bowman said AI revenue was growing quickly but was not yet sufficient to cover the cost of maintaining and developing models and data centers. Outside the Fed, former SEC Chair Gary Gensler told CNN he expects a bubble-driven recession before the 2028 election, while economist Ed Yardeni argued that falling valuation multiples for chipmakers point to a market already correcting rather than a repeat of 1999–2000.

Waller's comments may reassure some investors, but they do little to settle the broader debate over the sustainability of the AI-driven borrowing boom. As the ECB noted, cheaper models could pressure today's leaders, and the key question is not just whether AI proves useful, but whether the companies making the largest commitments earn enough, soon enough, to pay for them.