Federal Reserve officials expressed widespread concern in October about the potential financial risks if the current artificial intelligence (AI) boom fails to meet investor expectations, according to minutes released Wednesday from their latest policy meeting. The discussion highlighted growing vulnerabilities tied to rapid borrowing and complex financing arrangements connected to AI-related investments.
The Fed's committee meeting opened with an unusual focus on systemic risks, emphasizing that debt linked to AI projects has expanded quickly, often through "complex and opaque financing structures." Several officials noted that these exposures were becoming concentrated within a relatively small group of firms, increasing the likelihood that any significant downturn could have ripple effects across the broader financial system.
AI-driven investments have become a key growth engine for the U.S. economy, as technology companies borrow heavily to build data centers and acquire specialized semiconductor chips. However, Fed officials voiced apprehension that if projected returns on these investments falter, the resulting losses could spread to the lenders and investors backing these efforts.
Concerns highlighted in the minutes followed a wave of investor skepticism on Wall Street during the weeks leading up to the meeting. Two prominent AI companies reported revenue growth that fell short of forecasts, while a major tech firm involved in expanding data center capacity unveiled financing plans that exceeded investors' expectations. Notably, Google’s market capitalization dropped by approximately $500 billion in a single day during October amid concerns that the company’s AI payoff timeline was longer than anticipated. Around the same time, Meta announced plans to raise up to $95 billion through a new stock offering, and Anthropic, known for its Claude chatbot, initiated what is expected to become the largest initial public offering in history.
Fed staff highlighted particular risks stemming from so-called "circular financing arrangements" in which chipmakers and other AI-related suppliers also act as investors or lenders to their customers. These arrangements, coupled with long-term lease commitments that are not reflected as debt on company balance sheets, add layers of financial complexity that could obscure true exposure.
Although Chairman Kevin Warsh did not directly address AI-related debt concerns in his post-meeting remarks, he acknowledged that credit conditions had tightened for some businesses and households after the September rate hikes and observed investors becoming more selective. At a separate conference on AI’s impact on jobs and productivity held later in the week, Warsh outlined a scenario under consideration wherein AI underperforms, stock prices decline, and economic growth slows.
The minutes also reflected ongoing caution about the banking sector’s resilience. While several officials described the banking system as well capitalized, they warned that vulnerabilities could emerge quickly, especially given banks’ potential exposure to private credit funds and other nonbank lenders. In response, some Fed members advocated for stronger capital requirements to ensure that banks maintain adequate buffers to absorb losses and continue lending during periods of financial stress.
