The current investment boom around artificial intelligence could slow down sharply if expected returns from the technology turn out to be lower than forecasts, the Bank for International Settlements (BIS) warns in a new annual economic report.
Optimism around AI has become one of the factors that have supported the global economy and financial markets in recent months. Investments in semiconductors, data centers and energy infrastructure have increased, and large technology companies have become one of the main drivers of capital spending.
However, it is the scale and speed of this process that creates new risks. The BIS compares the current AI boom with previous technology investment cycles - the canal building of the 1830s, the railway boom in Great Britain in the 1840s, electrification in the late 1920s and the dotcom boom of the late 1990s.
A common feature of these periods was that a real technological breakthrough attracted more capital than it could justify the commercial return. As a result, such booms ended in investment reversals and downturns that affected not only individual companies but also the economy as a whole.
The five largest hyperscalers—large providers of cloud and computing power—could spend more than $1 trillion in AI-related capital investment in 2025-2026, according to the BIS. Those liabilities are already outpacing companies' profits and free cash flow, which is why some financing is being raised through debt markets.
BIS Managing Director Pablo Hernandez de Cos, quoted by The Wall Street Journal, said competition for future market share could have led to overinvestment in AI. If the technology does not deliver the expected returns, the sector could be vulnerable and the current investment boom could come to a screeching halt.
BIS is particularly concerned not only about the volume of investment, but also how it is financed. This is the debt of hyperscalers, AI labs, equipment suppliers, construction and engineering contractors involved in building data centers, energy facilities and other infrastructure.
If the largest customers slow or stop capital spending, companies in the supply chain could face a sharp drop in revenue and difficulties in servicing debt. This could increase pressure on fixed income, corporate bond and private credit markets.
BIS stresses that this is not to deny the long-term potential of artificial intelligence. On the contrary, AI can increase productivity in the economy, but the transition to this model is accompanied by uncertainty: it is not yet clear how quickly companies will be able to implement the technology in real production processes and monetize it at a scale sufficient to justify the current investment.
Another risk is related to infrastructure limitations. The development of AI already faces shortages of electricity, network equipment, powerful semiconductors and the ability to quickly build data centers. These constraints could drive up project costs and at the same time increase overinvestment if companies seek to secure capacity in advance through long-term contracts.
For the financial system, the key threat is that a possible disappointment in the returns from AI could coincide with already existing vulnerabilities — high valuations of technology companies, rising debt, activity in the non-banking financial sector and increased sensitivity of markets to changes in interest rate expectations. rates.
In this scenario, the AI boom can turn from a source of growth into a factor of financial instability. First, there may be a reduction in capital investment and a fall in orders in the infrastructure chain, then a deterioration in the position of borrowers, a widening of credit spreads and a revaluation of AI-related assets.
For businesses and investors, the main conclusion of the BIS is that technological progress does not abolish financial discipline. Even if artificial intelligence does become one of the key technologies of the coming decade, this does not guarantee that all current investments will pay off and that all companies participating in the race will remain sustainable.
BIS, or the Bank for International Settlements, is an international financial organization headquartered in Basel, Switzerland. The bank was established in 1930 and is considered one of the oldest international financial institutions. Its primary mission today is to promote global monetary and financial stability through central bank cooperation, research, statistics, analytical reports, and support of international financial standards. The BIS is often called the "bank of central banks" because it works primarily with central banks and international financial organizations.
