Artificial intelligence ( AI ) and trade are no longer separate policy domains – AI growth depends on globalized supply chains for semiconductors, computing infrastructure and digital services, while trade is increasingly shaped by who controls AI infrastructure, data flows and cloud capacity, according to a recent report.

Export volumes of AI-enabling goods have doubled from US$1.9 trillion in 2014 to US$3.8 trillion in 2025, finds the Allianz Trade research report, accounting for 15% of global trade and far outpacing the 40% growth in goods trade overall.

Asia dominates the supply side, the report notes, accounting for 65% of global AI-related exports and seven of the top 10 exporters, while the US has tripled its AI-related imports since 2023, underpinned by 5,427 operational data centres, good for 45% globally.

AI boom built in Asia

Over the past decade, global trade in AI-related goods has doubled, Allianz Trade points out, far outpacing the growth of overall goods trade and of non-AI-related goods in particular. This expansion has seen three distinct acceleration phases:

• The first acceleration occurred in 2017–2018 ( +15%  year on year and +11% YoY respectively ), driven by fundamental developments in need for AI-related goods and the introduction of the transformative architecture ( such as most large language models ).
• A second surge followed in 2021 ( +30% YoY ), driven by the release of early large-scale models, including GPT-3 ( 2020 ) and DALL·E ( 2021 ), and paving the way for the emergence of generative AI in 2022.
• More recently, advances in multimodal and agentic AI systems have reignited investment momentum. While AI-related goods exports reached new heights in 2025 ( +22% YoY ), January and February 2026 data show that the boom continues, especially in economies such as Taiwan ( +62% YoY ) and South Korea ( +105% YoY ).

Asia has firmly established itself as the centre of global trade in AI-enabling goods, the report shares, led by China ( 18% share of global trade ), Taiwan ( 12% ) and Hong Kong ( 11% ) on the podium, joined by Singapore ( 7% ), South Korea ( 6% ), Malaysia ( 4% ) and Japan ( 3% ) in the top 10.

Emerging players include Mexico, which recorded the fastest growth in 2025 at +62% YoY, while Thailand and the Philippines are also expanding as alternative hubs. 

Middle East could send chip prices surging

Over the past two years, two-thirds of the expansion in Asian semiconductor exports, the report details, were due to price increases and one-third due to higher volumes. Risks of energy shortages from the Middle East crisis could send prices even higher, given the already tight and extremely concentrated supply.

Taiwan, for example, operates the biggest foundry worldwide, the report notes, and occupies 70% market share in 2025. In Q1 2026, wafer shipments grew +28%, while revenue in US dollar terms expanded by +40.6%.

Apart from a structural shift in product mix, the sustained increase in semiconductor pricing, the report points out, is best understood as a demand-led phenomenon, with supply-side constraints serving as an amplifying mechanism.

The fact that the majority of AI-enabling goods exports are driven by a small group of countries, the report adds, also makes their trade model highly vulnerable to potential shocks.

Taiwan and Hong Kong are the most exposed to an AI bubble burst, with 74% and 59% of their exports respectively being related to AI goods. They are followed by Singapore and the Philippines ( both at 47% ), Malaysia ( 43% ) and South Korea ( 32% ).

In contrast, trade in AI-enabling goods accounts for just 15% of US exports, indicating that American exports are relatively well diversified.

Demand concentrated in US, partially Asia

Since 2023, Taiwan and the US, the report shares, have recorded the largest increases in AI-related imports, reflecting their pivotal positions in the AI supply chain: Taiwan through the import of capital equipment for semiconductor manufacturing, and the US through demand linked to its dominance in AI services and data centre infrastructure.

Since the mass-market deployment of large language models since 2023, the US has tripled its imports of advanced AI-related products, Allianz Trade says, reflecting massive domestic investment in AI and continued reliance on foreign semiconductor supply.

In Asia, beyond Taiwan, the strongest import growth has been seen in South Korea, reflecting its dual role in semiconductor production and consumers of advanced AI components, and Hong Kong, highlighting its important roles in semiconductor production and technology flows into China due to US export controls.

Overall, the AI semiconductor market, the report states, remains highly concentrated geographically: the US being the dominant end market for the most advanced generation of AI semiconductors, the EU and China as the other two large, but structurally distinct demand pools.