How one bank is managing the risk that AI spending dries up

• Key insight: Regions isn't lending directly to data center developers, but its executives acknowledged the bigger risk lies with borrowers economically tied to the AI ecosystem. They said the Alabama-based bank is capping growth in its AI exposure until it has more data on how durable the boom will be.
• What's at stake: According to a recent Bank of America survey, 34% of global fund managers identified AI hyperscaler capital spending as the most likely source of a future systemic credit event.
• Expert quote: "Soundness, profitability, and growth, in that order, matters — especially when you're thinking about industries like this where there could be change." — Anil Chadha, Regions Chief Financial Officer

Are banks prepared for the credit consequences of a slowdown in the AI spending boom? It's an important question that has sprung up quickly, as AI-related capital spending has come to dominate U.S. economic growth.

On Friday, executives at Birmingham, Alabama-based Regions Financial were asked about the bank's exposure and how they're preparing for what happens when the music stops. They said Regions is treating a possible curtailment in AI spending like any other concentration risk on the bank's books.

The bank is having "discussions on a routine basis" to understand its portfolio composition and "connectivity," Regions CEO John Turner told analysts. He added that Regions is conducting "stress analysis" to answer the question of how it would be affected by "weakness" in the AI sector.

"Part of that is fundamentally just embedded in our concentration risk management analysis and the conversations that we have about that generally," Turner said. "As we think about portfolios, we think about credit risk. We're having ongoing conversations about the connectedness of exposure, interconnectedness of exposure, throughout that sector."

Gartner projects that total AI spending will hit $2.6 trillion worldwide in 2026. Meanwhile, IDC is forecasting that spending on AI infrastructure alone will reach $487 billion by the end of the year.

In May, 34% of global fund managers named AI hyperscaler capital spending as the most likely source of a future systemic credit event — double the share from just a month earlier — according to a Bank of America survey.

Earlier this week, Goldman Sachs CEO David Solomon told analysts on the bank's second-quarter earnings call that the AI tech boom is likely to "ebb and flow." There could be "bumps and recalibrations" in the next six to 18 months, Solomon said, adding that there is "a lot of uncertainty" around how tech infrastructure will be built and financed.

"Ultimately, you will have a recalibration, a reset, a drawdown and then a further acceleration," Solomon said during the company's July 14 call. "That's what the path generally looks like."

While data-center lending is mostly happening outside of the traditional banking system, banks such as Regions could still be exposed.

On Friday, RBC Capital Markets analyst Gerard Cassidy asked executives at the $161 billion-asset bank whether they're watching for a future credit problem tied to the possible bursting of an AI bubble. He said that he was not "suggesting AI is going to be a bust," but likened the recent AI boom to the dot-com build-out in the late 1990s and early 2000s — which, after demand failed to materialize, left internet companies and their lenders stuck with billions in bad debt.

While Regions has not directly financed data-center construction, Cassidy inquired about the bank's indirect exposure to borrowers that are economically dependent on the health of AI-driven spending.

Regions Chief Financial Officer Anil Chadha said the bank has an ethos of applying a "discipline of being cautious." He emphasized that Regions is deliberately not rushing to "get too far ahead of ourselves" in growing lending exposure tied to the AI ecosystem until its executives "gather the data."

"Soundness, profitability and growth, in that order, matters — especially when you're thinking about industries like this where there could be change," Chadha said.