AI boom lifts trade but raises financial risks – UNCTAD

AI-related goods are physical technology products connected to building or running AI systems. The article specifically includes advanced chips and servers used in data centres. These goods form part of the hardware foundation for AI computing and model development, rather than being AI software or services themselves.

Advanced chips perform the intensive computing needed for AI applications. Data-centre servers contain and connect this computing equipment at large scale. Together, they support the data-centre infrastructure that has become a major focus of investment in 2026. The article links this spending to the broader AI boom and to rising trade in technology-related goods.

Their importance reaches beyond the technology industry. AI-related goods represented about one-sixth of global goods trade by value in 2025 and generated 42 per cent of its growth. This concentration means that changes in demand for chips or servers could affect global merchandise trade and investment.

AI-related goods accounted for about one-sixth of global goods trade by value in 2025. That is a large share for a relatively focused group of products, including chips and data-centre servers. Their role was even greater when measured by growth rather than total trade.

The article says these goods contributed 42 per cent of the growth in global goods trade during 2025. This means their expansion supplied nearly half of the increase, despite representing only about one-sixth of the total value. The figures show how strongly AI hardware influenced the direction of merchandise trade.

The result is a trade boom with a narrow foundation. Global trade in goods and services grew 4.4 per cent in real terms in 2025 and was projected to grow about four per cent in 2026. UNCTAD warned that weaker AI hardware demand could remove a major source of momentum.

Global trade growth has become dependent on AI-related products because demand has expanded rapidly for the hardware behind AI. The article identifies chips and data-centre servers as central examples. These products are part of a technology-heavy investment cycle focused on advanced computing and model development.

The mechanism is concentration. AI-related goods accounted for about one-sixth of global goods trade by value in 2025, yet supplied 42 per cent of its growth. Investment is also being directed toward data centres, advanced chips and strategic technology projects. Higher prices have contributed to much of the recorded increase in trade values since March 2026.

This creates a narrow base for an otherwise broad global trade figure. Goods and services trade grew 4.4 per cent in real terms in 2025 and was projected to grow about four per cent in 2026. If AI demand weakens, other sectors may not immediately replace that momentum.

If demand for AI hardware slows, global merchandise trade could lose one of its strongest recent growth engines. UNCTAD specifically warned that weaker demand for AI-related hardware could remove a major source of momentum. The vulnerability comes from trade growth relying heavily on a limited group of technology products.

The scale is clear in the 2025 figures. AI-related goods, including chips and data-centre servers, represented about one-sixth of global goods trade by value. Yet they contributed 42 per cent of its growth. A decline in orders, production or prices for these goods could therefore weaken the increase in merchandise trade.

The risk also sits alongside higher prices, which have driven much of the recorded growth in trade values since March 2026. Global goods and services trade grew 4.4 per cent in real terms in 2025 and was projected to expand about four per cent in 2026. A hardware slowdown could make that outlook harder to sustain.

The financial risk comes from the close connection between AI investment, technology valuations and credit. UNCTAD said the AI boom shows signs of earlier financial bubbles. If technology share prices fall, investors may sell assets quickly, while lenders and markets become more cautious about extending finance.

The key mechanism is a feedback loop. Falling prices can trigger further selling, which puts additional pressure on technology companies and investors. A credit squeeze could then make borrowing more difficult or expensive for businesses outside technology. The trade body warned that this stress could spread well beyond the technology sector.

The article does not quantify the possible losses or identify a specific trigger. It does show why the exposure matters now: investment is concentrated in advanced chips, model development and data centres, while AI-related goods provided 42 per cent of goods-trade growth in 2025. A reversal could affect markets, credit and trade together.

UNCTAD identifies China and the United States as the main destinations for AI investment. Europe benefits to a lesser extent. Brazil, India, Indonesia, Kenya, Malaysia, Mexico and Thailand also benefit, but by varying degrees. The pattern shows that AI investment is distributed unevenly across countries.

The leading position of China and the United States is linked in the article to concentrated spending on advanced chips, model development and data centres. Investment flows are increasingly shaped by industrial policies, technological competition and geoeconomic fragmentation. These forces direct capital toward strategic technology projects rather than mainly toward short-term profitability or the cost of capital.

The United States illustrates the scale. Data-centre investment is expected to reach $400 billion in 2026 and approximately $600 billion in 2027. That compares with projected fixed-investment growth of seven per cent across all US sectors in 2026. UNCTAD says the investment picture would look profoundly different without data centres.

Foreign direct investment, or FDI, is investment by businesses or investors in productive activity located in another country. It matters because it connects capital with factories, infrastructure, technology and other long-term projects. The article reports that global FDI rebounded to $1.6 trillion in 2025, after two years of decline.

Capital is now moving toward strategic projects such as AI computing infrastructure, data centres and critical minerals. These projects require substantial resources and can influence where future economic activity develops. The article says investors are focusing less primarily on short-term profitability or the cost of capital when directing funds toward these areas.

The shift is uneven. FDI inflows to developed economies rose 11 per cent in 2025, compared with two per cent in developing economies. UNCTAD links the pattern to geoeconomic fragmentation, industrial policies and technological competition. Concentrated investment may strengthen leading technology hubs while giving other countries more limited benefits.