A Federal Reserve official is questioning if AI is becoming 'too big to fail.'

      A senior official at the Federal Reserve has raised an uncomfortable issue. This week, Kansas City Fed president Jeff Schmid mentioned that the financial aspects surrounding the AI expansion now require close attention. He observed that the industry has grown sufficiently for policymakers to start considering it from a macroeconomic perspective and questioned whether AI is evolving into a sector that is too big to fail.

      This phrase carries significant implications, as it echoes the discussions during the 2008 financial crisis, when banks had become so integral that governments felt compelled to bail them out. Its application to AI marks a notable shift in the official discourse.

      Schmid's concern revolves around the scale and interconnectedness within the industry. He proposed that the current surge in AI investment should be compared to previous economic booms, cautioning that concentration at this magnitude could transform troubles within a single sector into broader economic issues.

      The figures behind this anxiety are astonishing. Major technology companies are currently responsible for nearly $2.4 trillion in AI spending commitments, a sum that eclipses most prior corporate investment periods, leaving little margin for error should demand falter.

      The issues lie not just in the volume but also in the financing methods. The Bank for International Settlements has cautioned that a downturn in AI could impact credit markets similarly to the 2008 crisis, primarily because much of the growth has been financed through debt and reciprocal arrangements.

      These interconnected financial structures are what concern regulators. When chip manufacturers, cloud service providers, and model developers invest in each other, any setback at a single point can have widespread repercussions, and this exposure can be difficult to gauge externally.

      Markets have started to pay attention. Nvidia's substantial $750 billion in AI deals has pushed its credit default swaps to unprecedented levels, indicating that even lenders to one of the sector's most robust companies are acknowledging increased risk.

      While comparisons to past economic bubbles exist, they are not entirely accurate. By certain metrics, the AI boom resembles the dot-com era, with valuations and concentration exceeding levels seen in 2000, although today's leading firms are generating genuine profits unlike many of the popular companies in 1999.

      This presents a dilemma for the Fed. The real earnings give the impression that the boom is more solid than a speculative frenzy, yet the scale and leverage of the spending could allow even a thriving sector to transmit shocks if market sentiment shifts.

      Schmid's statements also relate to monetary policy. If AI investments continue to drive demand for electricity, semiconductors, and construction, it complicates the central bank's assessment of inflation and its decisions regarding interest rates.

      This spending is also putting pressure on the companies themselves. Big Tech's expenditures on AI are nearing their free cash flow, compelling firms that previously financed everything internally to increasingly rely on debt and external capital.

      This shift is what draws the Fed into the narrative. As financing transitions from corporate finances to credit markets and private lenders, the risks extend to institutions that the central bank is responsible for monitoring.

      Schmid is not alone in expressing concern. An increasing number of officials and analysts are beginning to liken the AI cycle to previous economic manias, even though few are willing to declare the peak of a boom that continues to generate substantial revenue.

      The primary concern is over concentration. A small number of enormous companies now make up a significant portion of market gains and capital expenditures, meaning any misstep by one of them would have repercussions well beyond the technology industry.

      However, none of this constitutes a prediction of a collapse. A Fed president contemplating systemic risk is fulfilling his responsibilities, and raising a concern early is intended to avert future problems rather than to forecast them.

      Nonetheless, the terminology is significant. When officials start using the phrase too big to fail, they indicate that AI has transitioned from being merely a market issue to being a matter of the overall system's stability.