The head of the Bank for International Settlements warned on September 10 that the scale of artificial-intelligence investment is creating new questions for financial stability. BIS estimates cited by Reuters put expected AI spending by the five largest global technology companies at more than $1 trillion across 2025 and 2026, with global AI investment potentially reaching $4 trillion by 2030.
The concern is not that AI investment is automatically a bubble. It is that large capital requirements, opaque financing structures, high valuations and concentration in a small number of firms can transmit shocks through markets if expected returns disappoint. Debt-heavy funding can amplify that transmission because lenders and bond investors are exposed alongside shareholders.
For central banks, AI also complicates the economic signal itself. Productivity gains could raise potential growth, while job displacement, investment booms and asset-price movements can change inflation and demand in different directions. The policy challenge is therefore to distinguish durable productivity improvement from financial excess without choking off useful innovation.