China’s export resilience stems from broad manufacturing advantages, while AI is currently boosting prices rather than actual export volumes

Overview

The report analyzes the true sources of China’s accelerating exports in 2026, distinguishes the price effects generated by AI from actual export volume growth, and examines final demand in emerging markets, the “new new three,” imports and the trade surplus, and risks from EU trade policy. Its central conclusion is that the resilience of Chinese exports continues to stem primarily from broad-based manufacturing competitiveness rather than AI products themselves; external friction may create pressure but is not yet sufficient to alter the constructive export outlook.

Core views

China’s nominal exports accelerated significantly in 2026, growing 18.5% year over year through July, the fastest pace since the post-pandemic reopening in 2021; export volumes still rose 10.8%, or approximately 11%, in the first half. The report believes that the contribution of prices to nominal growth has increased, but strong real export volumes show that the cost and product competitiveness of Chinese manufacturing remains the deeper source of support. Based on broad manufacturing advantages and emerging-market demand, Goldman Sachs raises its forecast for 2026 export volume growth to 7.9%, implying 5.3% year-over-year growth in the second half, and expects growth of approximately 5% over the following several years. AI-related trade has provided a substantial boost to nominal exports, but its contribution to real growth and domestic value added is limited. The report adopts a narrow definition covering only semiconductors, printed circuits, storage and processing units, and computer components, while excluding downstream electronic end products. Under this definition, AI-related products contributed nearly half of nominal export growth in the second quarter of 2026 and made incremental exports more concentrated in certain ASEAN markets involved in assembly, testing, and packaging, as well as in Northeast Asian technology-producing economies. However, this boost came mainly from prices rather than quantities: rapid data-center expansion lifted hardware demand, and prices for some memory chips have risen to five to seven times their mid-2025 levels; AI-related products contributed around two-thirds of China’s recent export price growth, but their export volumes declined 1.4% year over year in the second quarter. The gross value of AI exports also overstates their contribution to value added in the Chinese economy. Customs data record the gross value of goods crossing borders rather than the value added generated within China; semiconductor supply chains span multiple economies, and although exports from foreign-owned factories in China are counted as Chinese exports, part of their value may accrue overseas. Approximately 80% of nominal AI-related exports are completed through processing trade or special customs supervision zones, typically involving imported components that are assembled or integrated in China and then re-exported. Consequently, AI-related goods currently have a limited net effect on the trade balance, and their contribution to domestic value added is far below what gross export values suggest; their contribution to real export growth is also less than 1 percentage point and has weakened as recent volume growth has slowed. After stripping out price effects, real export growth comes from a broader range of high-tech manufacturing industries. The report uses a product-level Laspeyres volume index that fixes export values at the previous year’s level and applies the volume growth rate of each product to distinguish price changes from genuine changes in export volumes. The results show that electrical equipment and machinery excluding AI products remain the largest individual sources of real export growth. The “new three”—solar modules or cells, electric vehicles, and lithium-ion batteries—account for only about 5% of nominal exports but have recently contributed approximately 1.5 to 2 percentage points to real export growth, approaching the contribution of the “old three,” comprising furniture, home appliances, and apparel. The contribution of chemicals has also risen recently because supply disruptions in the Middle East prompted global buyers to turn to Chinese producers. By destination, ASEAN, the EU, and other emerging markets in Africa, the Middle East, and Latin America are the main sources of incremental growth. China’s real exports of electrical machinery and transportation equipment to the EU grew 19% year over year in the first half of 2026. The report further distinguishes the destination to which goods are shipped from the region where Chinese value added is ultimately absorbed to determine whether exports to emerging markets reflect local demand or transshipment. Except for ASEAN and Mexico, the gap between these two measures is small across most emerging markets: these economies receive approximately 39% of the gross value of Chinese goods exports while absorbing approximately 38% of the Chinese value added embodied in foreign final demand; consumption accounts for about 60% of local final demand, indicating that household demand rather than mere transshipment is the primary driver. ASEAN remains the most important transshipment exception, receiving approximately 12% of Chinese goods exports but absorbing only around 9% of Chinese value added in final demand. Nevertheless, even in ASEAN, approximately three-quarters of Chinese exports still correspond to local demand, particularly for electric vehicles and capital goods. Mexico receives 3.3% of Chinese goods exports and absorbs 2.2% of Chinese value added in final demand; the report finds no evidence that it experienced large-scale transshipment during the 2025 trade friction comparable to that in 2018, possibly because Mexico raised tariffs on China and strengthened enforcement under US pressure. The United States absorbs more Chinese value added than the value of its direct imports, showing that indirect supply chains still exist, although related transshipment is now more concentrated in ASEAN than in the past. From 2019 to 2024, the share of Chinese value added absorbed by emerging markets rose by approximately 7 percentage points, with consumption contributing around 6 percentage points, further supporting the view that local final demand is strengthening. China’s officially designated “new new three”—AI, robotics, and innovative drugs—currently make only a very small quantifiable contribution to total exports. AI services are difficult to capture fully in official services trade data, and many Chinese models are free or priced very cheaply, so usage may be high while cross-border revenue remains limited; OpenRouter data show that Chinese suppliers accounted for 71% of global token consumption in July 2026, but the platform is oriented toward individual users and provides inadequate coverage of enterprise demand and commercialization. China became a net exporter of industrial robots in 2025, and robot exports totaled US$747 million in the first half of 2026, but accounted for only 0.03% of total exports during the period. Overseas licensing transactions for innovative drugs grew 66% year over year to approximately US$100 billion in the first half of 2026, but transaction payments generally depend on milestones such as clinical trials and regulatory approvals and do not immediately translate into an equivalent amount of export revenue; total year-to-date services exports during the same period were US$218 billion. Therefore, the near-term significance of the “new new three” is more strategic in nature. They are more oriented toward services and intellectual property than earlier export industries and can enable China to export AI capabilities, automation solutions, and pharmaceutical licensing revenue, while helping it move up the global value chain and broaden its export markets and product mix. However, their statistical coverage is incomplete and commercial revenue is realized slowly, so they are not yet sufficient to materially change the overall export trajectory; whether they can generate meaningful incremental growth over the long term depends on sustained support for high-tech industries and actual monetization. On the import side, growth in 2026 was also driven mainly by prices and concentrated in a small number of categories. Nominal imports grew 26.6% year over year in the first half, while import volumes increased only 7.8%; the second-quarter increase in real imports came mainly from gold and did not represent a broad recovery in domestic demand, with Goldman Sachs’s domestic demand tracker remaining at a low level of approximately 2 percentage points in recent months. Energy import volumes were a significant drag on second-quarter performance, possibly reflecting both high price sensitivity and ample domestic inventories; if global energy prices stabilize and inventories are gradually drawn down, energy import volumes could recover moderately. Gold imports have increased sharply since the beginning of 2026, but the report believes this partly reflects support from quota policies rather than market demand alone. China’s imports of nonmonetary gold require quota approval, while exports are generally restricted to a small number of authorized institutions, so actual import volumes depend on how the authorities adjust quotas. AI-related imports mainly serve supply chains and processing trade rather than final domestic demand, a characteristic that has become increasingly evident under policies promoting high-tech self-reliance and supply-chain security. After removing the distortion caused by large gold imports, China’s underlying trade surplus widened to an annualized level of approximately US$1.3 trillion through July. Given weak domestic demand, Goldman Sachs expects import volumes to grow 4.4% in 2026 before slowing to approximately 2.5% thereafter. External policy risks have shifted from the sharp US-China tariff increases of 2025 toward more managed and selective friction. The effective US tariff rate on Chinese imports was broadly stable at 23% in 2026; the temporary Section 122 tariff expired on July 24 but was offset by new Section 301 tariffs related to forced labor. China’s effective tariff rate on US imports is 27.4%, still higher than the US rate on Chinese products, while strategic competition over AI, advanced technology, investment, and critical minerals continues. The more direct downside risk comes from the EU. The EU accounted for 15% of China’s exports in 2025, and Chinese shipments to Europe have recently accelerated; advanced manufactured goods such as electrical machinery, lithium-ion batteries, vehicles—particularly electric vehicles—and computers are also areas where the China-EU trade surplus is concentrated. The report expects EU policy to become more assertive but to remain focused on industries facing clear trade diversion or industrial pressure, including steel, plug-in hybrid electric vehicles, electrical machinery, wind-turbine components, and basic chemicals. Existing and proposed measures could marginally cover approximately 27% of China’s annual nominal exports to the EU, but coverage does not equal actual export losses, and the ultimate impact will depend on policy details and implementation. Major new policies include expected countervailing duties of 17% to 38% on plug-in hybrid electric vehicles and an expansion of the Carbon Border Adjustment Mechanism to downstream products. The latter could expand coverage from basic steel and aluminum products—which affected only 2.1% of China’s exports to the EU in 2025—to an additional US$58 billion of exports, raising the affected share to 12.1%. The additional coverage primarily involves electrical equipment, machinery, and transportation equipment; these industries contributed 4.9 percentage points to the 8.5% growth in China’s nominal exports to the EU in 2025. However, the downstream expansion is not expected to take effect until 2028 at the earliest, and charges are based on the embodied carbon emissions of steel and aluminum precursors in finished products rather than the full value of those products. The effective tax burden may therefore represent only a small portion of the final export price, although high-carbon products such as heavy trucks would be more significantly affected. The report believes that broader European tariffs would threaten China’s gains in European market share, but the actual impact could be cushioned by several factors. China’s cost advantages and improving product quality remain supportive; even though battery electric vehicles have faced the EU’s 10% basic import tariff and manufacturer-specific countervailing duties since October 2024, they recovered rapidly in 2026 after an initial decline. Exports of internal-combustion-engine vehicles and plug-in hybrid electric vehicles, which are not subject to the additional battery electric vehicle tariffs, grew rapidly, showing that exporters can adjust their product mix and adapt to policy changes. China supplies more than 90% of the EU’s rare-earth elements by weight, retaining leverage over critical inputs; European businesses also have incentives to preserve access to the Chinese market and avoid forceful retaliation. Goldman Sachs therefore expects EU policies to tighten but possibly avoid measures that would provoke a severe response from Beijing, leaving the ultimate impact on the growth of China’s exports to the EU moderate overall.

Analysis framework

The report first defines AI-related goods using a narrow measure, then decomposes nominal exports into price and volume contributions and uses the share of processing trade and the trade balance to assess their domestic value-added content. It subsequently applies a product-level Laspeyres volume index to identify the industries driving real exports and uses the Asian Development Bank’s multiregional input-output tables to compare shipment destinations with the regions where Chinese value added is ultimately absorbed, thereby distinguishing local demand from transshipment. Finally, the report combines gold-adjusted imports and the trade surplus, industry-level export data, and specific tariff-policy scenarios to formulate its export forecast and assess the impact of China-EU trade friction.

Methodology notes

• Industry/Sector Analysis FrameworkPrice-volume decomposition

  Product-level Laspeyres export volume index

  The report fixes product export values at the previous year’s levels and then applies the volume growth rate of each product, thereby removing price changes and decomposing nominal export growth into price effects and genuine changes in export volumes.

• Industry/Sector Analysis Framework

  Final-demand absorption analysis using multiregional input-output tables

  The report uses the Asian Development Bank’s multiregional input-output tables to compare the direct shipment destinations of Chinese goods with the regions where the Chinese value added embodied in them is ultimately absorbed, thereby identifying local emerging-market demand and transshipment trade.

• Industry/Sector Analysis Framework

  Narrow proxy measure for AI-related trade

  Because AI trade lacks a unified definition, the report includes only semiconductors, printed circuits, storage and processing units, and computer components, while excluding downstream electronic end products, in order to isolate as much as possible the trade changes caused by expanding AI computing capacity rather than general digitalization.

Key data

• Nominal export growth through July 202618.5% year over yearThe fastest growth since the post-pandemic reopening in 2021
• Export volume growth in the first half of 202610.8% year over yearApproximately 11%, showing that real exports remained strong
• Contribution of AI-related products to second-quarter nominal export growthNearly halfDriven mainly by prices rather than quantities
• Second-quarter export volume growth for AI-related products-1.4% year over yearRecent volume growth has weakened
• Contribution of AI-related products to recent export price growthApproximately two-thirdsPrices for some memory chips have risen to five to seven times their mid-2025 levels
• Share of processing trade and special customs supervision trade in AI-related exportsApproximately 80%This means the contributions to domestic value added and the trade balance are lower than gross export values suggest
• Share of nominal exports accounted for by the “new three”Approximately 5%Including solar products, electric vehicles, and lithium-ion batteries
• Contribution of the “new three” to recent real export growthApproximately 1.5 to 2 percentage pointsClose to the contribution of the “old three”
• Growth in China’s real exports of electrical machinery and transportation equipment to the EUUp 19% year over year in the first half of 2026Compared with the first half of 2025
• Chinese goods exports received and Chinese value added absorbed by emerging marketsApproximately 39% and approximately 38%The two figures are close, showing that most exports correspond to local final demand
• Change in the share of Chinese value added absorbed by emerging markets from 2019 to 2024Up approximately 7 percentage pointsConsumption contributed approximately 6 percentage points
• Robot exports in the first half of 2026US$747 millionOnly 0.03% of total exports
• Overseas licensing transactions for innovative drugs in the first half of 2026Approximately US$100 billionUp 66% year over year, but revenue is generally realized gradually as milestones are achieved
• Import growth in the first half of 2026Nominal growth of 26.6%, volume growth of 7.8%Growth was driven mainly by prices and gold imports
• Underlying trade surplus excluding goldApproximately US$1.3 trillionAnnualized through July 2026
• 2026 export volume growth forecast7.9% year over yearImplies 5.3% growth in the second half, followed by expected growth of approximately 5% over the next several years
• 2026 import volume growth forecast4.4% year over yearExpected to slow to approximately 2.5% thereafter
• Effective US-China tariff rates23% on Chinese imports into the US, 27.4% on US imports into ChinaThe effective US tariff rate on China was broadly stable in 2026
• Potential coverage of EU measuresApproximately 27% of China’s annual exports to the EUCoverage does not equal actual export losses
• Proposed downstream CBAM coverageAn additional US$58 billion, raising the total affected share to 12.1%Expected to take effect in 2028 at the earliest
• China’s share of the EU’s rare-earth supplyMore than 90% by weightProvides China with leverage over critical inputs in trade-policy negotiations

Impact & implications

The report implies that the foundation of China’s export growth is broader than AI-related nominal data suggest: electrical equipment, machinery, automobiles, and other high-tech manufacturing industries, together with genuine final demand from emerging markets, can continue to support medium-term export volume growth. AI and the “new new three” have long-term strategic value, but price effects, processing trade, and delayed revenue realization currently limit their contribution to domestic value added. Meanwhile, weak import demand will keep the underlying trade surplus elevated; tighter EU policies may affect exports from some advanced manufacturing industries, but cost advantages, the ability to adjust products, leverage over critical materials, and European commercial interests are expected to cushion the overall impact.

Risks

• If the EU expands trade restrictions targeting steel, automobiles, electrical machinery, wind-turbine components, and basic chemicals, it could weaken China’s gains in European market share.
• Proposed countervailing duties on plug-in hybrid electric vehicles and an expansion of the downstream CBAM could raise the costs of relevant Chinese export products, with the actual impact depending on final policies and implementation details.
• Strategic competition between China and the United States in AI, advanced technology, investment, and critical minerals could continue to pressure trade and supply chains.
• A further escalation in protectionism and geopolitical tensions could weaken the resilience generated by the diversification of China’s export markets and products.

What to watch

• Monitor the industries, tariff rates, and implementation timetable ultimately targeted by the EU’s trade-diversion monitoring framework.
• Monitor whether countervailing duties on plug-in hybrid electric vehicles fall within the expected range of 17% to 38%, and whether the downstream CBAM is implemented as early as 2028.
• Monitor whether AI-related exports can shift from price-driven growth to volume growth and increase the shares of ordinary trade and domestic value added.
• Monitor ASEAN’s transshipment share and whether consumer demand in emerging markets can continue to support Chinese exports.
• Monitor whether milestone payments for innovative-drug licensing, commercialization of AI services, and robot exports can enable the “new new three” to generate meaningful revenue.
• Monitor the impact of gold import quotas, energy prices, and changes in domestic inventories on import volumes and the underlying trade surplus.
• Monitor how effectively Chinese exporters respond to European trade policies through product-mix adjustments, cost advantages, and quality improvements.