BIS warns the AI investment boom could also become a financial risk

The Bank for International Settlements says global investment tied to artificial intelligence helped support recent growth, but also introduces a new source of vulnerability. Its 2026 annual report does not dismiss AI’s productivity potential. It warns that the speed, concentration and financing of the buildout could amplify risks if expectations are not met.

What happened

The Bank for International Settlements, or BIS, published its Annual Economic Report 2026 on June 28, alongside a speech by General Manager Pablo Hernández de Cos at the institution’s annual meeting in Basel. The core message mixes optimism and caution: the global economy proved resilient in 2025, partly because of AI-related investment and sentiment, but fiscal and financial foundations are under greater pressure.

In the official speech, the BIS lists among the vulnerabilities the risk of an abrupt end to the AI investment boom, stretched financial valuations, high public debt and the rise of non-bank financial intermediaries. CNBC summarized the report by saying debt, financial fragilities and the sustainability of the AI boom are raising global risks. Reuters also reported the BIS warning from a global markets angle.

Why it matters

The signal matters because it comes from an institution that serves as a forum for central banks, not from a technology company or an investor with a direct stake in the AI cycle. The BIS recognizes that spending on data centers, chips, power, equipment and software has been a real source of demand. According to the report, AI-related investment softened part of the impact of tariffs and uncertainty and, in the United States, may have contributed roughly one percentage point to real GDP growth in 2025, based on estimates cited by the institution.

But the report also presents the other side. The five largest hyperscalers are on track to spend more than $1 trillion on AI-related capital expenditure from 2025 through 2026. Those bets may be rational if demand and productivity arrive as expected. If they do not, excess capacity, debt and valuations could become a stress channel for credit markets, equities and broader financial conditions.

What changes for companies and the AI ecosystem

For companies adopting AI, the warning does not mean every project should stop. It does suggest a sharper separation between verified productivity and market narrative. A pilot that reduces time, errors or costs has a different logic from investment driven only by fear of falling behind.

For infrastructure providers, the report points to concrete bottlenecks: electricity, advanced semiconductors, construction and grid equipment. If the race to secure future capacity is financed through long contracts, debt or circular arrangements among labs, suppliers and major platforms, a revenue disappointment could spread faster.

Social and strategic context

Lía Torres’ reading is that AI is no longer only a product discussion. When a technology mobilizes capital, power, debt and productivity expectations at macroeconomic scale, it also becomes a stability issue. That does not make the AI boom an inevitable bubble, but it does require looking beyond the lab: electricity prices, provider concentration, supply chains, fund exposure and labor effects.

What remains unclear

The BIS does not prove that the AI cycle will end in crisis or predict which companies will win or lose. Its scenarios depend on future productivity, real demand, infrastructure costs and financial conditions. Not all AI spending is the same either: frontier model training, inference operations, internal process modernization and speculative hardware capacity carry different risk profiles.

The responsible conclusion is narrower: the AI investment boom is now large enough to appear on the macro-financial risk map. If productivity follows, it can support growth. If it disappoints, it can amplify vulnerabilities that already exist.

Sources consulted

BIS speech: Read More Annual Economic Report 2026: Read More Read More Read More by Lía Torres — Social and strategic perspective.