Kansas City Fed’s Schmid asks whether the AI ecosystem is becoming too big to fail

Jeff Schmid, president of the Federal Reserve Bank of Kansas City, said the central bank needs to understand whether the artificial-intelligence “ecosystem” is getting too big to fail¹. He raised the question on Friday². Asked about AI, Schmid questioned whether the ecosystem is becoming “too big to fail”⁷, and wondered aloud whether artificial intelligence and data centres are creating an ecosystem that fits that description, arguing that the Fed needs to understand it⁵.

The Economic Times framed the remarks as the AI boom raising too-big-to-fail concerns as the ecosystem expands³. Schmid separately said US debt seems “extreme”⁶.

He is not the only official attaching a macro consequence to the build-out. The Cleveland Fed’s president has said AI spending is fuelling inflation⁸, and has said rate rises may be needed to tame it⁸. Federal Reserve Governor Michael Barr said on Wednesday that further interest-rate rises are likely needed⁶.

Why it matters to investors

“Too big to fail” is not a neutral phrase inside a central bank. It is the category the Fed uses for institutions whose failure it cannot allow, and which it therefore supervises and capitalises differently. A regional Fed president applying it to a capital-spending cycle, rather than to a bank, signals that AI funding is now being examined as a financial-stability question and not only as a growth story^1,3.

The bond market is already treating it that way. Investors are turning cautious on corporate bonds related to AI⁵. The New York Times editorial page put the same question directly, noting the difficulty of reconciling a too-big-to-fail framing with the message coming out of the leading AI companies themselves⁴.

Bill Ackman has warned that the Fed may have made a “mistake” and that AI could break its inflation model⁸ – an argument that cuts in the opposite direction from Schmid’s, and one worth holding alongside it rather than instead of it.

What to watch

The first thing to watch is whether the framing spreads from speeches into supervisory language, because that is the step that would change how banks and non-bank lenders are examined on their AI and data-centre exposures^1,2.

The second is the rate path. Barr has said further rises are likely needed⁶ and Cleveland has linked AI spending to inflation⁸, so the sector’s capital demands are now an input to policy rather than a sideshow to it. Whether Schmid’s question gets an answer from the Fed, or only more questions, is the thing to follow^3,5.