The Bank for International Settlements (BIS) is sounding an alarm that should make any finance leader sit up: an AI downturn could become a sharper, faster crash than a traditional banking crisis. BIS’s warning, delivered in its annual economic report released on Sunday, lands at the intersection of two things markets have been riding hard, AI investment and rapid capital movement through channels that do not look like classic banks.

Here is the core mechanism BIS highlights. As the global financial system struggles to keep pace with the artificial intelligence investment boom, funding for AI is increasingly being channelled through hedge funds and private credit, according to the BIS report referenced by SCMP. BIS also points to capital flows surging through “loosely regulated, non-bank channels.” Put differently, the money is moving, and it is moving in ways that may amplify stress when the cycle turns.

To understand why this matters, it helps to think about how traditional banking crises behave. Banks are regulated institutions, and stress often shows up in relatively familiar places: deposit runs, interbank lending freezes, and credit contractions tied to bank balance sheets. BIS is essentially saying that the AI boom changes the transmission path. When AI financing and related capital exposure build up inside less regulated non-bank vehicles, the stress can transmit differently. Non-bank channels can be more sensitive to market funding conditions and risk appetite shifts, and those shifts can happen quickly.

BIS’s framing is not just about where money sits today. It is about system resilience during a downturn. If AI investment momentum slows, the funding side does not simply pause. It can unwind. Expectations about growth can reverse. Valuations can compress. Liquidity can tighten. And because the BIS report connects this to “capital flows” surging through non-bank pathways, the worry is that the unwinding could accelerate, creating the kind of feedback loop that turns a slowdown into a crash.

This is especially relevant right now because AI has been a capital magnet. The AI investment boom has pulled in investors and lenders across the market spectrum, from equity-focused players to credit providers and asset managers. When the BIS report says the financial system is struggling to keep pace, the implication for boards and senior executives is that infrastructure, oversight, and risk management assumptions may not be aligned with how fast AI-related capital has been moving. The faster money moves, the more difficult it is for risk controls, stress testing, and governance to remain ahead of reality.

The “loosely regulated, non-bank channels” phrase is doing a lot of work. It signals a regulatory gap: the system may be relying on parts of the financial plumbing that are not constrained in the same way as banks. That does not mean every non-bank vehicle is reckless. It does mean the collective behavior in a downturn could be harder to predict and harder to contain. Hedge funds and private credit can play constructive roles in normal markets, but when liquidity dries up, their interactions with credit availability, collateral, and investor demand can magnify volatility.

BIS is also implicitly challenging a common comfort that decision-makers have historically taken from the banking-centric nature of past crises. If your mental model is “the banking system breaks, then the economy follows,” BIS is arguing that AI-related capital networks could produce a different, potentially more abrupt cascade. That is the “sharper, faster” part of the warning. The risk is not only the magnitude of losses, but the speed at which financing conditions can deteriorate across interconnected market segments.

So what should executives and boards do with this? They should treat the BIS warning as a governance checklist for AI exposure and funding assumptions, not as an abstract macro note. If your firm uses hedge-fund-style or private-credit-style counterparties, or if your business model depends on continued AI financing at favorable terms, your risk lens needs to cover second-order effects. In a downturn scenario, the question is not only “will demand drop?” It is “can liquidity retract faster than our planning horizon?” BIS’s report points to a world where the answer may be yes.