The Bank for International Settlements has issued an urgent warning about a dangerous new form of circular AI financing, a self-reinforcing web of equity stakes, supplier contracts, and debt that the central bank of central banks says could turn the artificial intelligence boom into a $1 trillion systemic risk. Released on June 28-29, 2026, the BIS Annual Economic Report 2026 places this risk alongside the Iran conflict and record sovereign debt as the top threats to global financial stability. AI capital expenditure drove roughly 74% of U.S. GDP growth in the first quarter of 2026, and the five largest hyperscalers — Alphabet, Amazon, Meta, Microsoft, and Oracle — are on pace to spend more than $1 trillion.
What Is Circular AI Financing?
Circular financing occurs when hyperscalers and chipmakers take equity stakes in AI labs that commit to purchasing computing power and chips from those same backers. In effect, money circulates within a closed loop: the investor provides capital, the AI lab books revenue by buying from the investor's own infrastructure, and the investor reports both equity upside and supplier income. According to the BIS Annual Economic Report 2026, these interlocking arrangements carry poorly disclosed terms and create a risk that the same underlying asset can be pledged multiple times.
Key mechanics include:
- Equity stakes from cloud providers and chipmakers into AI startups
- Take-or-pay compute contracts that channel revenue back to the investors
- Debt-financed capex that shifts funding from operating cash flow to leverage
- Off-balance-sheet shadow borrowing through private credit
How Circular Financing Echoes 2008 Rehypothecation
The BIS explicitly compares the structure to the rehypothecation risks of 2008, when the same collateral was reused to back multiple transactions, amplifying hidden leverage. Private credit loans to AI-related firms surged from about $3 billion in 2010 to more than $40 billion by 2025, and corporate bond issuance by hyperscalers topped $100 billion in 2025. If AI returns disappoint, a sudden withdrawal of financing could force fire sales across linked assets.
| Feature | 2008 Rehypothecation | 2026 Circular AI Financing |
|---|---|---|
| Core mechanism | Same collateral pledged to multiple counterparties | Equity, debt and supplier deals create self-referential revenue |
| Hidden leverage | Securities reused across repo chains | Off-balance-sheet private credit and take-or-pay contracts |
| Trigger | Mortgage losses and collateral calls | AI return disappointment and financing withdrawal |
| Transmission | Bank solvency crisis | Sovereign bond fire sales via leveraged hedge funds |
The Numbers Behind the Warning
AI capex has become a macroeconomic force. In the first quarter of 2026, AI-related spending accounted for roughly 74% of U.S. GDP growth, according to BIS calculations. The five largest hyperscalers deployed more than $1 trillion in AI capex across 2025 and 2026, and trillions of dollars in AI assets are set to begin depreciating between 2026 and 2028, compressing margins. Separate hyperscaler AI capex estimates point to even larger buildouts ahead.
- Amazon led 2026 guidance at about $200 billion, up from $131 billion in 2025.
- Alphabet guided to $180–190 billion; Microsoft effectively about $190 billion.
- Meta raised its range to $125–145 billion; Oracle is included in broader estimates.
From Tech Bust to Sovereign Debt Crisis
The BIS warns that the danger does not stop with technology. A new sovereign-financial stability nexus has emerged because highly leveraged hedge funds now dominate government bond markets. If an AI credit shock triggers margin calls and deleveraging, those funds could dump sovereign bonds, widening term premia and raising government financing costs just as public debt sits near record highs. This could transform a tech sector correction into a broader sovereign debt crisis.
What the BIS Is Saying
BIS General Manager Pablo Hernández de Cos delivered a blunt message: Policymakers must act now. The report urges central banks and regulators to strengthen prudential frameworks, demand greater transparency around circular financing, and rebuild fiscal buffers while financial conditions remain favorable. Waiting, the BIS argues, only raises the eventual adjustment cost.
FAQ: Circular AI Financing and Systemic Risk
What is circular AI financing?
It is a closed loop in which hyperscalers and chipmakers invest in AI labs that commit to buying from those same backers, inflating both revenue and asset values while hiding concentration and leverage.
Why does the BIS compare it to 2008?
Because the same AI-related assets can be pledged multiple times, similar to the rehypothecation that amplified losses during the 2008 financial crisis.
How much are hyperscalers spending on AI?
The five largest hyperscalers have deployed more than $1 trillion in AI capital expenditure across 2025 and 2026.
Could an AI bust cause a sovereign debt crisis?
Yes. Leveraged hedge funds now dominate sovereign bond markets, so an AI-driven margin call could force bond sales, raising government borrowing costs globally.
Conclusion
The BIS's 2026 warning is not a forecast of imminent collapse but a call to recognize that markets are not fully pricing the interconnected risk of circular AI financing. For investors and policymakers, the task is to separate genuine AI productivity gains from leverage-fueled circularity before the loop tightens.
Follow Discussion