Key Takeaways:
The Bank for International Settlements is pushing to complete Basel III Endgame rules that would make AI infrastructure financing significantly more expensive — threatening the $1 trillion-plus spending cycle underpinning the technology sector's biggest rally in decades.
The BIS used its latest Annual Economic Report to warn that the AI buildout has become a credit boom, not just a technology cycle, as companies from Nvidia Corp. to Microsoft Corp. borrow aggressively to fund well over $1 trillion in infrastructure spending.
"The current AI investment cycle depends on a continuous flow of capital that may not remain available under tighter regulatory standards," the BIS said in its report, which urged countries to complete implementation of the Basel III Endgame framework.
The rules would force banks to hold more capital against large corporate loans, restrict proprietary internal risk models, and impose stricter market-risk requirements under the Fundamental Review of the Trading Book. Corporate bond issuance has already surged as technology companies borrow faster than internal cash flows can support the buildout. If regulators succeed, the cost of financing everything from semiconductor fabrication plants to hyperscale data centers could rise sharply, potentially stalling a spending cycle that JPMorgan projects will reach $5.5 trillion by 2030.
The BIS report zeroes in on four structural risks embedded in the current financing model: concentrated AI investment, growing leverage across the technology sector, opaque financing arrangements between banks and borrowers, and expanding links between traditional banks and private credit markets. Each of these, the report argues, creates vulnerabilities that tighter regulation is designed to address.
Basel III's Impact on AI Financing
The Basel III Endgame represents the final phase of global banking reforms developed after the 2008 financial crisis. Banks would lose much of their ability to use proprietary internal models that often classify large corporate loans as relatively safe. Instead, regulators would require standardized risk calculations, stricter operational risk requirements, and higher capital charges for globally systemic banks.
Every change points in the same direction: banks would need to commit considerably more capital to support large, complex technology loans. That does not eliminate financing, but it makes it substantially more difficult and expensive. The last time regulators coordinated a tightening of this magnitude, during the initial Basel III rollout after 2008, banks tripled their capital reserves over a multiyear period.
Private Credit Isn't a Safety Valve
Many investors assume private credit funds can simply replace traditional bank lending if Basel III restricts bank financing. The BIS anticipates that argument and pushes back explicitly. Risk migrating from regulated banks into private credit does not reduce systemic risk — it merely hides it, the report argues.
The private credit sector has already shown cracks. Defaults are rising, payment-in-kind financing — which allows troubled borrowers to defer cash interest payments — is increasing, and redemption pressure from investors is growing. The sector has significant concentration in technology lending, making it particularly exposed to an AI spending slowdown. The BIS's long-term solution is to extend tougher oversight to private credit through leverage limits, enhanced reporting requirements, and stricter collateral standards.
What This Means for Investors
Current valuations across the AI supply chain assume years of uninterrupted capital spending and virtually unlimited access to financing. Goldman Sachs estimates that just four companies — Microsoft, Meta, Amazon and Alphabet — will account for roughly $5.3 trillion in spending between 2025 and 2030. Hyperscaler capital expenditures alone are expected to climb from $650 billion in 2026 to more than $1.1 trillion in 2027.
If regulators successfully restrict both bank lending and private credit while central banks keep interest rates elevated, they will not necessarily kill artificial intelligence. But they could dismantle the financial engine powering today's AI spending boom. Credit cycles have a long history of ending far more abruptly than technology cycles, and the current buildout depends on a continuous flow of capital that may not survive the regulatory squeeze.
This article is for informational purposes only and does not constitute investment advice.
