BIS warns: AI bubble could trigger global recession

What happened?

The Bank for International Settlements (BIS), known as the central bank of central banks, has published its 2026 annual report, dedicating a special chapter to artificial intelligence. According to the report, AI investment is reaching historic levels reminiscent of past bubbles: canals and railways in the 19th century, electrification in the 1920s, and the dot-com bubble of the 1990s. The BIS estimates that the five largest hyperscalers (Amazon, Microsoft, Google, Meta, and Oracle) will spend more than one trillion dollars on AI-related capex in 2026. This figure far exceeds the revenues and free cash flow of these companies, forcing them to issue debt to finance themselves.

The report notes that all past bubbles shared a common feature: a genuine technological breakthrough that attracted capital in excess of what commercial returns could justify. AI is no exception. Although companies report employee-level efficiency gains in pilots, few have achieved significant productivity gains in production. The BIS warns that fierce competition is leading companies to commit resources to projects with uncertain returns, which could reduce the industry's net economic surplus and even turn it negative in adverse scenarios.

Why is it important?

The BIS warning is crucial because it is not an isolated opinion but an analysis backed by the institution that coordinates global monetary policy. The scale of investment is unprecedented: Amazon alone projects $200 billion in capex for 2026, Microsoft $190 billion, Google $180 billion, and Meta up to $140 billion. If this bubble bursts, the consequences would not be limited to tech companies but would spread to the entire economy. The BIS identifies several risks:

• Supply bottlenecks: the availability of electricity, chips, and grid connections is already putting pressure on energy prices and input costs, with potential inflationary effects.
• Amplified overinvestment: companies sign long-term contracts to secure future capacity, exposing them to a demand downturn.
• Financial vulnerabilities: the high leverage of AI companies and their growing presence in credit markets could trigger a collapse in asset prices if optimistic sentiment shifts.
• Supplier ecosystem: engineering, procurement, and construction contractors have weak balance sheets and would suffer from capex cuts.
• Financial opacity: sector financing is opaque, with private agreements and data center lease terms not fully disclosed.

The BIS warns that if inflation picks up or AI investment collapses, macroeconomic consequences could be amplified by existing financial vulnerabilities, leading to a sharp adjustment in asset prices and disruptive macro-financial feedback loops.

What consequences will it have?

The consequences of an AI bubble burst would be severe and far-reaching. In the short term, an investment reversal would induce a global economic recession, similar to that following the dot-com bubble burst. Tech companies would suffer massive losses, but the impact would extend to suppliers, the financial sector, and the real economy. Central banks might be forced to raise interest rates to control inflation, worsening the asset downturn. Moreover, sector opacity could hide losses until it is too late.

In the long term, the lesson would be that AI investment must be more cautious and based on real returns, not expectations. The BIS suggests that financial regulation should pay more attention to exposure to the tech sector and transparency of financing agreements. For companies, the recommendation is to diversify and not bet everything on AI without a clear monetization strategy.

What should readers know?

Readers should understand that the AI bubble is not a certainty, but the risks are real and are being flagged by the highest global financial authority. It is not an imminent event but a warning for investors, companies, and regulators to act prudently. AI remains a transformative technology, but its massive deployment requires time and realism. The BIS's historical comparisons are revealing: in every past bubble, the underlying technology was genuine, but overinvestment led to losses and recessions. The question is not whether AI will change the world, but when and at what cost.

“All shared a common feature: a genuine technological breakthrough that attracted capital in excess of what commercial returns could justify.” — BIS Annual Report, 2026

For investors, the recommendation is caution: avoid overexposure to tech stocks and diversify. For companies, it is key to focus on use cases with clear returns and not get carried away by competition. For citizens, understand that the global economy is interconnected and a crisis in the tech sector can affect jobs, savings, and pensions.