Image Credit: Shuterstock
The Bank for International Settlements has warned that heavy AI investment could create financial stress if expected returns fall short of the money being spent.
The warning appeared in the BIS Annual Economic Report 2026, which listed AI capital expenditure alongside inflation, public debt and non-bank finance as pressure points for the world economy.
Five Hyperscalers May Spend $1T
The BIS said the five largest hyperscalers are expected to spend more than $1 trillion on AI-related capital expenditure across 2025 and 2026.
That spending includes semiconductors, data centers, computing capacity and power infrastructure. The report said those commitments are rising faster than earnings and free cash flow, pushing some firms toward debt financing.
AI Gains Could Still Take Years
The BIS did not dismiss AI’s potential value. It said the technology could lift productivity over the next decade, with task-level studies showing time savings of 20% to 50%.
The concern is timing and scale. If companies spend heavily before demand is proven, the build-out could shift from an investment boom into a longer pullback.
Chips and Power Limits Raise Costs
The BIS said AI infrastructure is already facing limits in electricity, advanced chips and grid equipment. Those constraints can raise costs and push firms to lock in long-term capacity through contracts before the market has settled.
That can leave companies exposed if demand, pricing or productivity gains fall short. The report compared the AI race with earlier technology booms, including railways, electrification and the dotcom cycle. Each involved a real technological breakthrough, but capital moved faster than commercial returns could support.
Debt Financing Raises Market Risk
For markets, the risk is leverage tied to data centers, chips and cloud capacity. The BIS said AI labs, hyperscalers and suppliers have issued large volumes of debt, while private financing arrangements remain hard to track.
A sharp reversal in AI optimism could hit equity values first, especially for companies priced on high long-term earnings growth. The BIS said US stocks now make up about 64% of the MSCI Global index, meaning a US-led equity selloff could move through global portfolios.
Credit Losses Could Spread to Suppliers
Credit markets are another weak point. If hyperscalers slow spending, suppliers tied to data center construction, chips and computing capacity could lose revenue and struggle to service debt.
The BIS also pointed to circular financing between chipmakers, hyperscalers, AI labs and cloud providers. These arrangements can help fund the boom while money is easy, but they can make losses harder to trace when funding tightens.
BIS Urges Non-Bank Finance Oversight
The BIS urged policymakers to strengthen oversight beyond banks, including private credit and other non-bank finance. It said supervisors need better visibility into AI-related financing, model-driven trading, correlated exposures and cyber risks linked to frontier AI systems.
The report is not a call to stop AI investment. The BIS is warning that policymakers need to track where debt, leverage and supply-chain exposure are building before a pullback in AI spending feeds into equity, credit or non-bank finance stress.
