Wall Street’s AI fear has a number now, and it’s 87%

A glowing AI monolith casting a shadow over a small bank building under storm clouds

The fear that the AI boom might blow up the financial system has gone from a fringe mutter to the single loudest worry on Wall Street, and now there is a number on it. In the Federal Reserve’s latest Financial Stability Report, released on 13 November, 87 per cent of the market contacts it surveyed named AI-related risk as a salient threat. A year ago that figure was 30 per cent. In the spring it was 50. The worry has nearly tripled in twelve months.

The numbers that matter

The headline figure is the trajectory: 30 per cent in autumn 2025, 50 per cent in spring 2026, 87 per cent now. But the detail underneath is worse. The report estimated that debt, guarantees and not-yet-commenced leases tied to AI infrastructure could exceed 3 trillion dollars. That is a far bigger number than the roughly 800 billion dollars of AI-related debt a Fed governor flagged only weeks ago, because it counts the promises and future commitments, not just the loans already drawn. The specific fears the contacts named were cyberattacks, the risks buried in complex AI-related financing, and the possibility that an AI earnings slowdown triggers a broad market correction. For context, the next most-cited risks trailed well behind: a risk-asset correction at 52 per cent, higher long-term rates at 48, cyberattacks at 43.

The comparison the Fed chose to make

Here is the part that should make you sit up. The report itself compared the current AI infrastructure build-out to the late-1990s broadband boom. That is the central bank, in its own words, reaching for the dot-com era as the nearest parallel. Anyone who remembers how the broadband build-out ended, enormous investment in capacity that took years to be used, and a brutal reckoning for the companies that borrowed to build it, knows that is not a flattering comparison. It is the Fed quietly saying it has seen this shape before.

The honest caveat

This is a survey of sentiment, not a forecast. The sample is small, 23 market contacts, and the report is careful to say the views are the contacts’, not the Board’s official position. So nobody at the Fed is predicting a crash. What the data captures is mood, and mood is exactly the thing that turns a wobble into a rout. A market where 87 per cent of serious participants are nervous about the same thing is a market primed to move hard the moment that thing twitches.

Bull, bear, neutral

Bull: widespread early awareness is the system working, not failing. Much of the 3 trillion is funding real infrastructure that will eventually earn its keep, and a worry this well-telegraphed rarely blindsides anyone.

Bear: near-unanimous nervousness is itself a fragility signal, because it means positioning is crowded and confidence is thin. Three trillion dollars of commitments against revenue that has not materialised is the classic set-up, and the Fed picked the broadband bust as its analogy for a reason.

Neutral: watch what actually sits inside that 3 trillion figure, whether private-credit stress surfaces, and above all whether the AI earnings slowdown the contacts fear shows up in real results. The numbers will tell you before the narrative does.

Not investment advice. Do your own research, and for anything serious, speak to a professional.

What this means

A year ago, calling AI a systemic risk to the financial system got you looked at funny. Now it is the most-cited worry in the Fed’s own survey, carrying a price tag that could top 3 trillion dollars, with the central bank itself reaching for a bubble as the comparison. None of that guarantees a bust. It does mean the people whose job is to watch for one have stopped whispering.

Edgar Friendly, calling the AI bullsh*t. More tools, more swearing: toptoolstack.com.

Did you know? The Fed’s survey has asked about AI for three years running. The share of market contacts calling it a financial-stability risk has gone 30 per cent, then 50, then 87. Few worries in the survey’s history have climbed that fast.

Sources: AOL / Reuters; MLQ; Superpower Daily.

Get the free weekly stack: the AI tools and moves that matter, hype filtered out.Subscribe free →
Scroll to Top