Federal Reserve Governor Michael Barr told attendees at a computing technology and finance conference in Milan on Monday that artificial intelligence poses significant risks to the U.S. financial system and global economy.
On par with previous investment booms fueling the expansion of railroads in the 1860s, the U.S. interstate highway system 100 years later and telecommunications infrastructure at the turn of the 21st century, so-called “advanced computing-related debt issuance” topped $680 billion year to date at the end of October.
“Each of these investment booms may have felt overbuilt at some point, but today we are glad to have the railroad, highway and fiber-optic capacity that they put in place,” said Barr, who oversees consumer protection and community development for the Fed.
“Given the still-limited amount of debt backing the current boom, I’m less concerned about direct losses,” he added, instead highlighting the network of impacts that could unfold during a large-scale repricing of AI-related assets.
AI risks dominate financial stability concerns
The U.S. central bank published its semiannual financial stability report Friday, which assesses risks to the U.S. financial system that may stem from “valuation pressures, borrowing by businesses and households, financial-sector leverage, and funding risks.”
The nearly 70-page report showed a sharp rise in concern from market-based survey respondents in financial vulnerabilities related to AI. Those concerns run the gamut from a price correction hitting overvalued technology stocks to a sudden sharp slowdown in AI-related spending.
Nearly 9 in 10 respondents cited AI as “presenting risks to the U.S. financial system and the broader global economy.” In the May edition of the survey, just 50% of respondents expressed those sentiments.
Barr suggested the financial security infrastructure may be unprepared for the collective risks of AI, asking the audience to imagine a bank run perpetrated and perpetuated by AI agents.
“Where customers once needed to execute transfers, AI agents could instantaneously detect yield differentials or stability signals and autonomously move funds at high speed, potentially accelerating runs that unfold in minutes, overwhelming traditional liquidity management and regulatory intervention timelines,” cautioned Barr.
Economists say that roughly half of U.S. economic output is currently attributable to AI spending and investments. But the sustainability of the AI boom remains uncertain, Barr emphasized, pointing to how the rapid growth in AI-related debt has raised concerns over circular financing and “complex financial obligations,” such as residual value guarantees that both secure current investments and underpin forward-looking valuations.
Get these articles in your inbox
Sign up for our daily newsletter
Get these articles in your inbox
Sign up for our daily newsletter
“With leverage at hedge funds and broader risk appetite apparently quite high,” said Barr, “a sharp reversal could lead to rapid deleveraging and, potentially, self-reinforcing fire sales.”
Equities market, corporate bonds fuel valuation fears
Barr has spoken repeatedly over the past year about financial vulnerabilities related to the build-out of AI and the longer-term economic transition such investments have precipitated. He has flagged limited engagement from peer regulators at the Consumer Financial Protection Bureau amid this transition as a particular cause for concern, especially as economic growth and broader financial stability have become increasingly reliant on the returns of AI spending.
The Fed’s financial stability report examines the largest asset classes of the U.S. economy, the top five of which are public equities, residential real estate, Treasury securities, commercial real estate and investment-grade corporate bonds.
Concerns of a risk asset correction emerged as the second most heavily mentioned concern in the report published Friday, with more than half of respondents citing that as a risk to financial stability and the broader economy, compared to one-third in the spring.
The public equities market was valued at $93.6 trillion at the end of the second quarter, up more than 25% from a year earlier as technology stocks related to the AI boom continued to push U.S. equity indexes higher.
By comparison, the $61.7 trillion residential real estate market reflected price growth of just 2.3% from a year ago. Just three years ago, residential real estate was the largest U.S. asset class at $56.3 trillion, followed by public equities at $53.4 trillion.
The size of the investment-grade corporate bond market — into which major technology firms have sold massive amounts of debt this year to fund a historic expansion in computing infrastructure to power AI — grew 7.8% year over year in the second quarter to $8.6 trillion.
In the fall of 2023, the investment-grade debt market was valued at slightly under $7.4 trillion, underscoring the massive valuation shifts fueled by AI enthusiasm but backed by investor cash.
“Advanced computing-related debt issuance rose to more than $680 billion by October 2026, about a fivefold increase from $130 billion in all of 2023,” said the report, noting that estimates of outstanding public debt related to advanced computing exceeded $1.1 trillion as of last month.





