BIS Warns $6.7T in AI Funding Risks Could Rival 2008 Crisis

The world’s central bank for central banks has just issued one of its starkest warnings about the financial system in years — and it isn’t about inflation or sovereign debt. It’s about AI. The Bank for International Settlements published its annual economic report on June 28, placing AI funding risks at the center of a potential global financial disruption that, in the worst case, could rival or exceed the damage caused by the 2008 crisis.

Key takeaways

• The BIS identifies AI as one of three major global “pressure points” alongside inflation and sovereign debt in its June 28 annual report.
• Most AI investment flows through hedge funds and non-bank financial intermediaries with limited regulatory oversight, concentrating systemic risk outside traditional banks.
• McKinsey estimates $6.7 trillion in cumulative capital expenditure will be needed by 2030 for AI and computing infrastructure — $5.2 trillion for AI-enabled data centers alone.
• BIS warns that any market correction could move faster than during previous banking crises, partly because households now hold higher equity exposure than in prior decades.
• The BIS urges immediate regulatory reform, cautioning that the longer oversight improvements are delayed, the more chaotic any subsequent market adjustment will be.

BIS Warns of AI Funding Risks Comparable to 2008

The concern isn’t that AI itself will collapse — it’s that the way AI is being financed looks dangerously familiar. A rapid concentration of investment, high leverage, limited transparency, and a dependence on markets that lack traditional regulatory guardrails: these were ingredients in previous boom-and-bust cycles. The BIS believes they are assembling again, this time around artificial intelligence.

The institution frames AI as one of three “pressure points” squeezing the global economy right now. The other two — persistently elevated inflation and record sovereign debt — are already well-understood threats. AI is the newer and, in some ways, more unpredictable variable. Unlike inflation or fiscal imbalances, AI investment risk is harder to measure, harder to stress-test, and largely invisible to regulators who focus on the traditional banking sector.

AI as a Key Economic Pressure Point

AI has boosted investor confidence and driven growth expectations tied to long-term productivity gains. But the BIS report makes clear that those expectations carry serious risks if returns don’t materialize. Supply bottlenecks and fierce competition among AI developers and cloud technology providers could generate the kind of excess investment that has historically ended badly.

The parallel the BIS draws is not subtle. The report places the current AI investment cycle alongside historical boom-and-bust patterns: British canal building in the 1830s, the railway boom of the 1840s, the electrification wave of the late 1920s, and the dot-com surge of the late 1990s. Each ended with a sharp downturn when investment outpaced the ability to generate sustainable returns.

Potential for a Rapid and Severe Market Correction

What makes this moment particularly acute, according to the BIS, is speed. Zhang Tao, the BIS chief representative for Asia and the Pacific, was explicit: “The speed of a correction could be much faster than previous banking crisis episodes.”

Part of that acceleration risk comes from the fact that households are now more exposed to equity markets than at any previous point in comparable cycles. A sharp decline in tech equity valuations — which AI spending has been directly propping up through earnings projections and corporate growth forecasts — would hit ordinary investors harder and faster than the correction of 2008, when the damage was initially concentrated in mortgage-linked instruments and only later spread to the broader economy.

Concentration of AI Funding in Non-Bank Financial Intermediaries

The structural problem is where the money is coming from. The bulk of AI investment does not flow through regulated banks — it moves through hedge funds, private credit vehicles, and non-bank financial intermediaries (NBFIs) that operate with far less regulatory scrutiny. This is not a marginal observation. It is the central vulnerability the BIS is flagging.

Role of Hedge Funds and Private Credit Vehicles

Since the 2008 financial crisis, banks have genuinely strengthened their balance sheets. They are better capitalized, and systemic leverage within traditional banking has come down. But that risk didn’t disappear — it migrated. Leverage and liquidity exposure shifted into investment funds, hedge funds, private equity vehicles, and other intermediaries that exist largely outside the regulatory perimeter designed to detect and contain systemic shocks.

BIS General Manager Pablo Hernández de Cos put it plainly, noting that high debt levels today are being financed through non-bank financial intermediaries — and the urgency for policymakers to address that cannot be overstated. “Policymakers must act now. Delay will only make the necessary adjustments more costly,” he said.

Risks in Shadow Banking and Regulatory Oversight Gaps

The so-called shadow banking sector has been a recurring concern in BIS research for years. What’s new is the direct link to AI investment. The financing chains supporting AI infrastructure — from data center construction to chip procurement to cloud capacity buildout — now run deeply through private credit markets and non-bank lenders. If investor sentiment turns, losses in these channels could amplify rapidly, and regulators may simply lack the visibility to intervene in time.

The BIS also flagged a newer dynamic: sovereign debt markets increasingly dominated by large, highly leveraged hedge funds have created what the institution calls “a new sovereign-financial stability nexus.” Any sharp repricing in sovereign bonds could tighten financial conditions globally, compounding any AI-driven disruption. Frank Smets, acting head of the BIS monetary and economic department, warned that such swings “could rapidly tighten financial conditions.”

Economic Consequences of an AI Investment Downturn

Projected Capital Expenditure Requirements and Investment Boom

The scale of what’s at stake makes the risk calculus even more serious. According to McKinsey estimates cited in the BIS report, the world will need approximately $6.7 trillion in cumulative capital expenditure by 2030 just to meet rising demand for compute power. That breaks down as roughly $5.2 trillion for AI-enabled data centers and $1.5 trillion for traditional IT infrastructure. Global data center capacity could nearly triple by 2030, with around 70% of that demand driven by AI workloads.

These are not speculative projections — they represent investment commitments already being made by corporations, woven into earnings guidance and long-term strategic plans. That integration is precisely what makes a potential correction so economically dangerous.

Risk of Prolonged Underinvestment and Market Volatility

If AI investments fail to generate the returns that justify these outlays, corporate decision-makers will pull back. Capital expenditure programs will be deferred. The investment boom could flip into a prolonged period of underinvestment, with knock-on effects on employment, growth, and the availability of financial capital more broadly. The BIS describes this not as a tail risk but as a plausible scenario that warrants serious preparation.

This is the deeper economic threat — not just a stock market correction, but a structural pullback in spending that could suppress productivity and growth for years, much as the dot-com bust left a hangover in technology investment well into the mid-2000s.

Calls for Transparency and Regulatory Reforms

The BIS is not simply raising alarms — it is calling for a specific policy response. The institution urges policymakers to extend and strengthen oversight beyond the traditional banking sector, improve transparency around non-bank funding sources, and prioritize price stability while pursuing fiscal sustainability. The emphasis on non-bank oversight is not incidental; it is the piece most urgently missing from the current regulatory architecture.

The report was published ahead of the European Central Bank’s annual Sintra symposium, where global policymakers are expected to debate many of the same stability concerns. For AI specifically, the BIS identifies three watchpoints: whether corporate earnings from AI investments justify the capital already deployed, whether that capital spending holds at current levels, and whether regulators succeed in improving transparency over the non-bank channels fueling AI’s growth.

The institution’s message on timing is unambiguous. The longer reforms are delayed, the more disorderly any adjustment will be. In a market where AI spending is now embedded in corporate budgets, equity valuations, and household wealth through retirement accounts and investment portfolios, an unmanaged correction would not stay contained. It would move — fast, broadly, and through channels that current oversight frameworks were not built to handle.

FAQ

Why does the BIS warn that AI funding risks could be worse than the 2008 financial crisis?

Because AI investments are highly concentrated among a limited number of investors and largely financed through non-bank entities with little regulatory oversight, any rapid deleveraging could trigger global market disruptions. Unlike 2008 — where risk was concentrated in bank balance sheets and mortgage instruments — AI risk sits in harder-to-monitor channels like hedge funds and private credit markets, and households now carry greater equity exposure than they did then, amplifying the potential economic fallout.

What makes the funding structure for AI particularly risky according to the BIS?

Most AI capital flows through hedge funds, private credit vehicles, and non-bank financial intermediaries that lack the regulatory oversight applied to traditional banks. These entities can amplify losses rapidly if market sentiment shifts, and regulators currently have limited visibility into the size and interconnectedness of these exposures.

What economic impact could a decline in AI investment returns have?

A meaningful drop in AI returns could lead corporations to cut or defer major capital expenditure programs, turning the current investment boom into a prolonged period of underinvestment. This would carry broader consequences for economic growth, employment, and the availability of financing across the technology sector.

What measures does the BIS recommend to address these risks?

The BIS calls for improved transparency around non-bank financial intermediaries, stronger oversight that extends beyond the traditional banking sector, and prompt fiscal and financial reforms. It warns explicitly that delays in implementing these measures will make any subsequent market adjustment significantly more chaotic and costly.

Article produced with the assistance of artificial intelligence and reviewed by the editorial team.