Jeff Schmid, president of the Federal Reserve Bank of Kansas City, said in a speech this week that investment into artificial intelligence (AI) has grown large and interconnected enough to require monitoring at the macroeconomic level. Schmid posed a pointed question: could the AI sector become “too big to fail,” a phrase with strong associations to the 2008 financial crisis?
Size, concentration and interconnections
Schmid’s worry centers on the sector’s scale and the tight financial links among its participants. He noted that major technology companies have collectively committed roughly 2,400 billion dollars in AI investments. According to Schmid, that volume exceeds past corporate investment cycles and leaves limited room for error if demand falls short of expectations.
Beyond raw investment totals, Schmid highlighted the financing structures. The Bank for International Settlements (BIS) has warned that a potential AI market downturn could severely shake credit markets because a large share of development is financed with loans and layered financing arrangements. Those interlocking business relationships mean the failure of a single participant could set off a chain reaction, while external observers may find actual exposures hard to trace.
Market signals and examples
Markets are beginning to price these risks. Schmid cited the wave of AI deals tied to Nvidia—reported at about 750 billion dollars in value—which has pushed up the cost of the company’s credit default swaps (CDS). That move signals that lenders to even the sector’s strongest firms see elevated risk.
At the same time, Schmid stressed the current environment is not identical to prior bubbles. Some indicators—valuation levels and market concentration—recall the dot-com era, but today’s leading tech firms are generating real profits, unlike many favorites from 1999–2000. This creates a dilemma for the Federal Reserve: profits suggest the expansion rests on firmer ground than pure speculation, but the sheer scale of investment and reliance on credit mean an otherwise healthy industry could still transmit financial shocks if sentiment turns.
Implications for monetary policy and supervision
Schmid warned that growing AI investment could complicate the Federal Reserve’s task. If the investment wave increases demand for energy, semiconductors and construction capacity, it can make inflation assessment and interest-rate decisions more difficult. Furthermore, AI spending is approaching the free cash flow capacity of tech giants, so firms that previously financed projects internally are increasingly relying on debt and outside capital.
That shift moves risk from corporate balance sheets toward credit markets and private lenders, exposing financial institutions whose supervision often falls within the Fed’s remit.
Conclusion
Schmid did not forecast a collapse. A regional Fed president’s role includes calling attention to potential systemic risks before they materialize. Yet using the phrase “too big to fail” in connection with AI signals a notable change in tone: AI is no longer only a market or technology story but has become a matter for financial-stability consideration.
How regulators respond—and how the sector’s financing evolves—will determine whether the benefits of AI investment can be reaped while containing the risks that arise from concentration and leverage.
