BIS warns AI spending boom may end in bust if returns lag
Source: Fintech News SG
The BIS general manager says funding AI's build-out with rising debt and opaque structures could backfire if the returns never arrive.
The Bank for International Settlements' top official has a blunt warning: if AI delivers underwhelming returns, the build-out funded by today's spending spree could flip from boom to bust. Pablo Hernández de Cos, the BIS general manager, told Mumbai's Global Fintech Fest 2026 that swelling debt loads and murky financing structures warrant close scrutiny for financial stability.
The scale is what makes it consequential. Five of the world's largest technology companies are on course to spend beyond US$1 trillion on AI across 2025 and 2026 combined, while industry participants see global AI investment climbing from roughly US$500 billion now to between US$3 trillion and US$4 trillion by 2030.
As that outlay outpaces the cash these firms generate internally, they lean harder on outside funding, and Hernández de Cos flagged opaque ties that make genuine value difficult to trace. Under circular financing, he noted, technology suppliers hold stakes in AI businesses and then sign up to buy their chips and computing capacity.
Fierce rivalry could also push companies to commit capital that future earnings cannot justify, he added. A market correction would bite household spending, since consumers now hold more of their wealth in equities, and the pain would travel worldwide because US stocks make up such a large share of global equity markets.
A second worry: AI sharpens both attackers and defenders in cybersecurity, and institutions must cover every gap while an intruder needs just one opening. The technology also muddies the signals central banks read on financial conditions, which feed into economic activity, inflation and interest rates. He wants better measurement, adaptable policymaking and greater international cooperation.
Why it matters for Singapore: Singapore sits on both sides of this trade, hosting regional operations for global banks, chipmakers and cloud players while courting AI investment as a core growth bet. A cooler AI capex cycle would quickly tighten the funding, listings and data-centre demand that flow through the city-state. Regulators here, long focused on AI governance, would need the same flexible, well-measured approach Hernández de Cos urges.
