BIS chief warns AI capex arms race relies on opaque debt, posing systemic risks

The escalating capital expenditure race among technology companies rushing to lead the artificial intelligence market is increasingly financed through opaque leverage, creating systemic vulnerabilities reminiscent of past speculative manias. Pablo Hernández de Cos, General Manager of the Bank for International Settlements (BIS), cautioned that the sheer velocity of capital deployment—driven more by competitive panic than by immediate profitability—could trigger severe macroeconomic fallout if commercial revenues fail to meet hyper-inflated expectations. Drawing direct parallels to the 19th-century railway boom and the dot-com crash of the late 1990s, Hernández emphasized that transformative technologies historically attract vast waves of speculative capital before financial realities set in. Today, tech behemoths and specialized startups alike are committing hundreds of billions of dollars to build massive data centers, acquire specialized graphics processing units, and secure long-term energy contracts. Crucially, a growing fraction of this spending relies on off-balance-sheet financing, private credit arrangements, and vendor financing models that mask the true extent of debt accumulation across the corporate landscape. This lack of transparency prevents regulators and credit agencies from accurately assessing the financial system's exposure. The core risk stems from a widening disconnect between extreme capital intensity and realistic monetization pathways. While infrastructure costs compound rapidly, enterprise adoption of generative AI has produced relatively modest revenue streams so far. Should the anticipated productivity gains lag behind schedule, the capacity to service these opaque debt structures will deteriorate. Given the deep integration between technology balance sheets, private debt funds, and institutional credit networks, a sharp repricing of AI infrastructure assets would not stay confined to Silicon Valley. Instead, a default wave or asset devaluation could rapidly propagate through non-bank financial intermediaries, tightening credit conditions for the broader economy. Furthermore, this financial engineering extends into overlap zones with digital asset markets, where compute power and data center yields are increasingly packaged into synthetic yield products and speculative collateral. Hernández urged global financial authorities to establish stricter disclosure requirements for technology-focused debt instruments and intensify surveillance over non-bank leverage. Without clearer visibility into how these capital expenditures are ultimately funded, central banks face an escalating risk: a technological breakthrough engineered into an unmanageable systemic crisis, where a forced market correction precipitates an economy-wide deleveraging shock.