Global trade is “riding” the AI boom, but dependence on the AI cycle is also increasing

Overview

This report discusses how the AI boom is driving global trade and the potential macro impact if the AI cycle cools. HSBC points out that despite headwinds such as tariff-policy volatility, Middle East shipping disruptions, and trade frictions, global merchandise trade has remained resilient over the past 18 months, mainly because of the rapid growth in trade in AI-related goods. At the same time, non-tech export growth has broadly stagnated since 2024, making global trade growth more dependent on the AI cycle.

Core views

The report’s core view is that AI-related goods are dominating global trade growth, but that also means the trade outlook is highly tied to the AI investment cycle. Asia is the biggest beneficiary of AI exports, with Taiwan and South Korea supported by hardware demand, while high-tech exports from Mainland China have also grown significantly. The US, Europe, and Hong Kong, among others, have higher shares on the import side. If the AI investment boom eventually reverses, trade growth may cool; however, HSBC does not currently treat this scenario as its base case, because hyperscaler capex is still expected to remain high and tariffs on AI-related products are generally low.

Analysis framework

The report uses a combination of macro trade shares, regional export structures, import dependence, capex forecasts, and scenario analysis. It first uses the WTO definition of AI-enabled goods to measure their share of global trade and contribution to growth, then examines concentration on the Asian export side and the US and European import side, and finally uses a static scenario in which export growth in Taiwan and South Korea is halved to estimate the drag on global export growth.

Methodology notes

• Macro trade analysisBreakdown of the trade contribution of AI-enabled goods

  Separate WTO-defined AI-enabled goods from other merchandise trade and compare their contribution to global export growth.

  This method is used to identify the portion of global trade growth driven by the AI value chain and to highlight the backdrop of weak growth in non-AI goods exports.

• Scenario analysisStatic scenario of halved export growth

  Assume that export growth forecasts for Taiwan and South Korea in 2026 and 2027 are halved, and estimate the drag on global export growth.

  This scenario is not HSBC’s base-case forecast, but is used to gauge the sensitivity of global trade growth to a cooling AI hardware cycle.

• Capex leading indicatorHyperscaler capex forecasts

  Use capex forecasts for the top six hyperscale cloud providers, including AWS, Google, Oracle, CoreWeave, Microsoft, and META, to judge the durability of AI hardware demand.

  Continued strength in capex implies that demand for AI-related components may continue to support trade, even if growth slows from extremely high levels.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Taiwan exports and the AI hardware supply chain

  Highly positively correlated

  Strengths

  About 80% of total exports are linked to the AI value chain, directly supported by AI hardware demand.

  Weaknesses

  The export structure is highly dependent on the AI cycle; if demand cools, trade growth will be hit more clearly.

  Comparison

  Compared with most economies, Taiwan has higher exposure to AI-related exports.

  Risks

  Reversal in AI investment, falling component prices, rising non-tariff measures.

• South Korea exports

  Positively correlated

  Strengths

  As an important participant in the Asian AI hardware chain, it benefits from AI demand.

  Weaknesses

  The report’s scenario analysis shows that if export forecasts for South Korea and Taiwan are halved, global export growth would be dragged down.

  Comparison

  Like Taiwan, South Korea is a beneficiary of AI hardware exports, but the growth forecasts disclosed in the report are lower than Taiwan’s.

  Risks

  A slowdown in the AI hardware cycle, global technology capex falling short of expectations.

• Mainland China high-tech exports

  Positively correlated

  Strengths

  High-tech exports grew nearly 40% YoY in 1H26 and account for 18% of global AI-related exports.

  Weaknesses

  If broader export engines slow, the aggregate impact could be larger given Mainland China’s high weight in global trade.

  Comparison

  Mainland China is both an important exporter and importer of AI-related goods.

  Risks

  Slower external demand, trade restrictions, rising non-tariff measures.

• US imports and data-center capex

  Demand-side support

  Strengths

  The US accounts for about 19% of global AI-related imports, about 40% of data centers are located in the US, and hyperscaler capex supports component demand.

  Weaknesses

  Price contributed a large share of nominal growth in US AI imports, while volume growth was relatively modest.

  Comparison

  The US is one of the most important import markets for AI-related goods.

  Risks

  Slowing capex growth, falling component prices, policy and tariff uncertainty.

• Digitally delivered services trade

  Long-term beneficiary

  Strengths

  AI may drive growth in digitally delivered services exports by improving productivity and lowering trade costs.

  Weaknesses

  Long-term forecasts depend on model assumptions, and the degree of short-term realization still needs to be observed.

  Comparison

  Compared with traditional services, digitally delivered services are considered by the WTO model to be the area where AI may drive the fastest growth.

  Risks

  Regulatory restrictions, cross-border data rules, cooling in the AI investment cycle.

Key data

• Share of AI-enabled goods in global merchandise tradeNearly 20%Higher than the average of about 14% in 2024.
• Contribution of AI-related goods to global merchandise trade growth in 2025More than 40%The report cites the WTO methodology.
• Contribution of AI-related goods to nominal YoY export growth in 1Q26About 80%Estimated by HSBC, which also notes that rising component prices boosted nominal value.
• Combined global AI-related export share of Mainland China, Taiwan, and Hong Kong44%By value in 1Q26; of which Mainland China 18%, Taiwan 14%, and Hong Kong 12%.
• Combined global AI-related import share of the US, Mainland China, and Hong KongAbout 49%US 19%, Mainland China 18%, Hong Kong 12%.
• Share of AI-related goods in Taiwan’s total exportsAbout 80%Shows that the export structure is highly dependent on the AI value chain.
• Share of AI-related goods in US imports27%Nearly doubled since the start of 2025, mainly driven by rising component prices.
• Change in US AI import value and volumeValue up 60% YoY, volume up 8% YoYIn the first five months of 2026, indicating that price effects had a large impact on nominal trade value.
• Scale of digitally delivered services exportsUSD 5.3 trillion, accounting for 55% of global services exportsDigitally delivered services exports grew 10% YoY in 2025.
• WTO estimate of AI’s long-term impact on digitally delivered services exports39%-42% growth in 2025-2040The report says this area may be most affected by AI-driven productivity gains and lower trade costs.
• Taiwan export growth forecastAbout 23% in 2026, 12% in 2027The report cites forecasts from HSBC economists for Taiwan and South Korea, on a real basis.
• South Korea export growth forecast7.8% in 2026, 3.1% in 2027The report cites HSBC forecasts, on a real basis.
• Global impact if Taiwan and South Korea export growth is halvedWeakens global export growth by about 0.2-0.3 percentage points per yearA non-base-case static scenario estimate.
• Capex forecast for the top six hyperscale cloud providersCould exceed USD 1 trillion in 2027YoY growth is expected to slow to about 25% in 2027, but the absolute scale remains high.
• Average tariff on AI-related goodsBelow 2% in high-income OECD economies, about 9% in least developed economiesThe report also notes that non-tariff measures are rising.

Impact & implications

For investment and macro judgments, the AI trade boom on one hand supports Asian exports, global merchandise trade, and digital services trade, while on the other hand it increases global trade’s sensitivity to a small number of products, a small number of regions, and the capex cycle of large technology companies. If AI capex remains strong, trade resilience may continue; if AI hardware demand or investment expectations reverse, global export growth, Asian manufacturing momentum, and related services trade could all come under pressure.

Risks

• The AI investment boom could ultimately reverse, like historical investment booms such as canal mania and the internet bubble.
• Non-tech export growth has broadly stagnated since 2024, showing a weak trade foundation excluding AI.
• Asian economies such as Taiwan and South Korea are highly dependent on AI hardware exports; if the AI cycle cools, export growth could slow significantly.
• If Mainland China’s export engine slows, the impact on overall trade growth could be larger because of its heavy weight in global trade.
• Nominal trade growth in AI-related goods has been driven by rising component prices; if prices fall, nominal trade growth could decline.
• Non-tariff measures on AI-enabled goods are rising and could weaken the trade support provided by low tariffs.
• Although capex remains strong, the growth rate is expected to slow, and the market may reassess the durability of AI demand.

What to watch

• Whether capex by the top six hyperscale cloud providers meets or falls short of expectations.
• Whether the divergence between AI-related component prices and actual import volumes widens or narrows.
• Whether high-tech export growth in Taiwan, South Korea, and Mainland China continues to outperform non-tech exports.
• Whether there is a turning point in the share of AI-related imports in the US and the pace of data-center investment.
• Whether digitally delivered services exports can maintain high growth and validate AI-driven improvements in productivity and trade costs.
• Whether tariffs and non-tariff measures on AI-enabled goods rise further.
• Whether global non-AI goods exports can emerge from the weakness seen since 2024.