Head of Bank for International Settlements Warns of AI Investment Bubble

The head of the Bank for International Settlements (BIS) has warned that the pace and scale of the current investment boom in artificial intelligence warrant caution, and that financing this spending with debt could threaten the stability of the global financial system. Pablo Hernandez de Cos made these remarks in a speech titled "Artificial Intelligence, Growth and Financial Stability Challenges: Central Banks" at a conference organized by India's central bank on September 10, 2026.

A Warning From Basel

Hernandez de Cos, who heads the institution sometimes called the "central bank of central banks," does not claim a crash is inevitable. He notes, however, that the scale of capital flowing into the AI sector is starting to resemble earlier speculative episodes in economic history, when investor enthusiasm outpaced actual returns on investment.

I do not say that this is where the AI boom must lead. But the scale and speed of the current investment boom, and the weight of expected commercial returns, do warrant some caution - Pablo Hernandez de Cos, head of the Bank for International Settlements

The BIS chief cited three historical analogies: the 19th-century canal and railway building frenzy, the electrification boom of the late 19th and early 20th centuries, and the dot-com bubble of nearly two decades ago. In all these cases, he noted, more capital flowed into the sector than the eventual returns on investment could justify, and the correction that followed had effects felt across the entire economy.

All drew in more capital than eventual returns could justify. In each of these cases, the eventual correction that followed had economy-wide implications - Pablo Hernandez de Cos, head of the Bank for International Settlements

Debt Instead of Profits

A key element of the warning concerns how the current boom is being financed. Capital spending by the largest tech companies on AI infrastructure is growing faster than their current earnings and cash flows, forcing them to turn to debt and private credit financing. Hernandez de Cos pointed out that this financing structure is far less transparent than classic equity financing, making it harder to assess the true scale of risk spread throughout the system.

Because U.S.-listed tech companies account for a large share of global stock market capitalization, any correction in valuations would not remain confined to the American market. The BIS chief stressed that the effects could spread globally, hitting emerging markets and countries closely tied to the U.S. financial system.

With U.S. stocks accounting for a large share of global equity markets, the effects could propagate globally - Pablo Hernandez de Cos, head of the Bank for International Settlements

Not Just Criticism

The speech was not one-sidedly pessimistic. Hernandez de Cos acknowledged that AI's promise as a source of productivity growth is real, citing research showing efficiency gains in specific professional tasks such as programming, consulting and professional writing ranging from 10 to 65 percent. He cautioned, however, that the technology's long-term impact on the economy depends on policy decisions, the quality of energy infrastructure, and whether productivity gains reach a broad range of workers or become concentrated among a narrow group of companies.

Another Warning in the Same Vein

This is not the first warning from the BIS this year. As early as June 2026, the institution published a report comparing the current situation to the dot-com bubble and warning about opaque financial transactions between AI giants, their lenders, and chip makers. The September speech expands on those findings with new figures and directly addresses the central bankers gathered at the conference in India, suggesting the topic will remain on the agenda of financial regulators worldwide.

For central banks, Hernandez de Cos noted, artificial intelligence does not formally change monetary policy mandates, but it significantly complicates reading economic signals, since it simultaneously affects demand, supply and financial markets. This means traditional tools for analyzing economic conditions may provide an incomplete picture in economies heavily exposed to AI investment.

What It Means for Investors

For Polish investors and companies indirectly involved in the AI supply chain, such as component makers or cloud service providers, the warning from Basel signals that global regulators are starting to treat the concentration of capital around a handful of U.S. tech giants as a systemic risk, not just a topic for industry debate. A potential slowdown in capital spending by hyperscalers could hit the entire chain of subcontractors, from server manufacturers to energy companies building data centers.