The investment boom in artificial intelligence, which supported the global economy last year, is now itself becoming a source of serious macro-financial risks. This is the conclusion reached by experts at the Bank for International Settlements in their latest annual report. Analysts point to issues such as debt financing of AI infrastructure, overvaluation of companies, rapid growth of private credit, and opaque deals between tech giants, chipmakers, and AI labs.

In 2025, the global economy successfully withstood tariff and geopolitical shocks, largely thanks to AI optimism, which stimulated capital expenditure and trade. However, this year the range of threats has widened. The regulator identified four key pressure points: the threat of persistent inflation, doubts about the return on AI investments, growing financial vulnerabilities, and deteriorating fiscal positions. Experts warn that euphoria could quickly turn to disappointment if production hits bottlenecks such as electricity shortages, advanced semiconductors, and network equipment.

The scale of investment is particularly concerning. Capital expenditures by the five largest technology corporations in 2025 and 2026 are set to exceed one trillion dollars. These outlays are already outpacing net income and free cash flow at some companies, forcing them to take on new debt. The BIS compares the current situation to historical technology bubbles: the canal mania of the 1830s, the British railway mania of the 1840s, the electrification euphoria of the late 1920s, and the dot‑com boom of the late 1990s. A common feature of these episodes was a genuine technological breakthrough that attracted more capital than commercial results could ultimately justify.

Opaque private deals within the AI sector are cited as a serious risk. So‑called circular financing is practiced: tech giants or chipmakers acquire stakes in AI labs, which in turn undertake multi‑year commitments to purchase computing capacity. The terms of such contracts are rarely fully disclosed, creating the risk of rehypothecation of the same assets. The regulator also notes that a correction in the U.S. market, where AI company stocks account for up to 64% of the global index, could quickly spread worldwide, hurting household wealth and global financial conditions.

An indicator of strain in the supply chain is the memory chip market. U.S. chipmaker Micron reported record quarterly revenue of $41.46 billion for the third quarter of fiscal 2026, having signed 16 strategic customer agreements. The remaining performance obligations (RPO) for already concluded agreements, including those signed after the end of the quarter, stand at about $100 billion, while the company also expects $22 billion in customer deposits and related financial commitments, of which around $18 billion will be cash deposits. Management acknowledges that memory supply is not yet keeping up with surging demand. DRAM brought in a record $31.3 billion, accounting for 76% of total revenue.

However, rising chip prices have already triggered legal action. On June 25, a class‑action antitrust lawsuit was filed in the Northern District of California against Samsung Electronics, Samsung Semiconductor, SK Hynix, SK Hynix America, and Micron Technology. The plaintiffs allege that the three largest DRAM producers coordinated supply curbs and price increases, and also shifted production toward more expensive AI products, including HBM memory. Earlier, in November 2025, media reported that Micron planned to invest $9.6 billion in memory chip production for AI in Japan.