Buy Now, Ban Later? How China’s Export Engine Is Fragmenting Global Supply Chains

Oliver Sawbridge

August 14, 2026 • 6 min read

---

This month, China’s trade surplus hit $126 billion. This is up $12 billion from a year earlier, making it the second largest surplus on record. In 2025, China’s goods-trade surplus was $1.19 trillion, which is bigger than the next ten countries put together. A surplus that size needs to go somewhere. 

Unsurprisingly, the US’s huge consumer market remains attractive, but with elevated trade barriers on products from China, goods are increasingly coming through countries where US tariffs are lowest. Just this week, the White House accused more than 40 jurisdictions, including Canada, Japan and the EU, of helping Beijing dodge US tariffs in what it’s dubbed “The Great Transhipment Scam.” “Made in China” could be turning into “Made Nowhere Else.”  

Washington’s likely response? Aided by new AI tools to help detect transshipment, we think it’s likely to turn the pressure outward: allies will be asked to police China’s trade themselves. In fact, we’re seeing it already with ongoing USMCA renegotiations, with the US pushing to block Chinese companies from reaching the US market via Mexico and Canada. 

Despite US and European efforts to curtail Chinese dominance, though, China’s value-add to global companies faced with intense cost pressures and the need for productivity-enhancing tech is arguably growing. As one global footwear manufacturer said to us: “We’re out of China… [but] I think 5% of our production is in China. The legacy of just about everything material is China, that’s where all of the technology is, that’s where the innovation is.”  

Now, operations and supply chain leaders will need to find the balance between over-exposure and premature withdrawal themselves, reconciling the need to decouple and derisk with relying on China’s strong supply chain capabilities.  

Open Hand, Closed Fist 

China’s ascent is not surprising: Beijing has been prioritizing its export machine as its property slump and weak consumer spending continues to weigh on growth prospects. But its manufacturing success is now provoking a backlash. Europe and America have moved first, with others likely to follow in an effort to protect domestic industry, with copious steel restrictions on China, for instance, already in place. 

This is intensified by the fact that China’s export playbook is changing, as demonstrated by its dominance in electric vehicles (EVs). Chinese brands have reached 9% of new car sales in the European Union and 15% in Britain in about two years. And these figures are only expected to rise.  

This is also no longer just a cheap-car story. Xiaomi is aiming at Europe’s premium segment and hiring from BMW, Porsche, and Tesla. Beijing is now doing the same with AI, most notably with the recent announcement of the China-created and UN-supported World Artificial Intelligence Cooperation Organization (WAICO). The organization offers developing countries cheaper AI systems and a voice in setting governance rules, through open knowledge sharing. 

The offer does sound good. Chinese firms pitch open-source models as a gift to countries and businesses that struggle to afford Western systems. Yet Beijing has been simultaneously discussing limits on foreign access to advanced models and tightening rules on cross-border AI deals. The underlying risk is structural: once Chinese models, chips, and data centers become the default, switching away means rebuilding around alternative hardware, software, and standards.  

A counterargument is that open-weight models can be downloaded once and run independently of Beijing. That may be true in theory. But China’s Kimi K3, for example, limits commercial use, and AI dependence sits well beyond the model itself – in chips, data centers, software, and technical support, which may prove much harder to replace in practice. 

China’s own compute constraints sharpen that logic. US export controls on advanced chips limit domestic capacity, giving Beijing more reason to seek influence over the overseas infrastructure that supports adoption of its models.      

Buyers Byte Back 

The EU’s trade deficit with China now runs at about €1 billion a day. That reflects a deeper mismatch: China produces 30% of the world’s manufacturing output but only 13% of its consumption. Such gaps lead to protectionist laws as governments attempt to shield factories at home, forcing companies to contend with more paperwork and regulatory complexity.  

Zero100’s analysis shows that before 2019, trade restrictions on China were sparse and consistent across every bloc. Since then, US-aligned countries have pulled sharply away on both measures. New restrictions peaked at nearly seven per country in 2025 – roughly 20x the 2019 level. China-aligned and neutral countries show no equivalent surge. 

The EU’s draft Industrial Accelerator Act would attack that gap directly by introducing local content requirements, ownership limits, and R&D spending rules. A proposed “overcapacity instrument” modeled on America’s Section 301, would go further, letting Brussels act against Chinese distortions without proving industry-wide harm first.  

And in Washington, there is movement to end duty-free exemptions for low-value parcels, an advancing Senate bill threatening Mercedes with a sales ban over Chinese supply-chain links, and anti-transshipment rules in the ongoing USMCA renegotiation to stop Chinese goods reaching the US market via Mexico and Canada.  

The same logic applies even more strongly to AI and to strategic technologies such as robotics, biotech, and telecoms. The problem for leaders is that policy has been erratic. Think President Trump’s reversal on the announced export restrictions on chips to China or the ruling that phones, computers and chips were exempt from tariffs. Without consistent government and political direction, companies must judge the risk for themselves. This is where the premature-withdrawal danger becomes concrete: some may hold back from China for political reasons, even in sectors where Chinese ecosystems are becoming central to global competition, including through de facto standard-setting.

A New Regulatory Reality 

Dependence inevitably brings intervention. Businesses building and using Chinese technology need to map that exposure now. Watch for procurement rules in AI-heavy public contracts, local content clauses in any “strategic sector” legislation modeled on the EU’s Industrial Accelerator Act, and outright bans following the Huawei precedent because of national security concerns – the same tools already used against cars and electronics. 

The lesson is to reassess exposure industry by industry. Chips, defense-related technology, data-heavy sectors, and consumer goods will not all be treated alike. Broad political assumptions will no longer do.  

As the divide between US and China-led blocs continues to widen, the safest assumption is that the regulatory burden will keep growing and the rules will keep shifting, especially as Washington and Beijing are not making common rules for AI and other emerging industries like quantum computing and biotechnologies.  

The task now is to balance: to build exposure mapping and contingency planning so as to take advantage of China’s manufacturing ecosystems, while limiting geopolitical exposure.