Currently, some officials addressing this issue advocate for caution, with a belief that a financial crisis similar to the housing collapse two decades ago—or the dot-com bubble, albeit to a lesser extent—is unlikely at this time. Nevertheless, the substantial levels of investment, unpredictable returns from an untested technology, and the emergence of complex financing arrangements and increased debt usage have placed AI finance on central bankers’ radar.
“I don’t perceive this as a bubble situation,” stated Federal Reserve Bank of New York President John Williams in a recent interview with Reuters.
“What we’re witnessing is an extraordinary degree of excitement and enthusiasm surrounding new technology, particularly AI,” Williams remarked. “Investors are actively attempting to tackle a nearly insurmountable question: how substantial will the benefits of AI ultimately be?” Addressing these inquiries is likely to induce volatility.
Williams noted that although borrowing to facilitate AI investment has risen, it is being managed by companies with significant earnings, adding, “I’m not overly concerned about financial stability from leverage at this moment.”
Torsten Slok, chief economist at money manager Apollo, highlighted in a research note, “The data-center expansion is still less than half the magnitude of the housing boom, which peaked at 6.6% of GDP in 2005,” while concurrently, the rate of investment relative to GDP is climbing faster than housing did prior to the global financial crisis.
MONITORING TIME
Some Fed officials appear to be less optimistic regarding AI risks compared to Williams.
Given the current financing methods in the industry, Kansas City Fed President Jeff Schmid stated in a Tuesday speech, “I would contend that there are indicators prompting us to address this on a macro level: is this industry becoming another too big to fail?” He specifically expressed concerns about financing flows and their potential to create vulnerabilities that may propagate issues through interconnected stages.
Schmid posed the question, “Is the cyclical nature of commitments—say, a contractual obligation to a data center from an energy provider and the community it serves—getting too leveraged? If a spark ignites a flame, what consequences might follow?”
San Francisco Fed chief Mary Daly remarked on Wednesday, “If you examine the growth rate and amount of investment in the AI sector, it would be easy to conclude that it’s quite concerning.”
To counterbalance that apprehension, Daly pointed out that many commitments within the AI realm are still in the announcement phase and have not yet materialized into physical investments, lowering the risk of “stranded assets” that could create difficulties post-shakeout.
However, she cautioned that the growing trend of borrowing to support expansion could present problems. For the Fed, “It’s really about assembling a dashboard focused not on past financial crisis triggers, but on potential risks that could destabilize the current situation.”
