Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Twenty-twenty-five was the year global trade definitely did not die. But it changed in complex ways, some temporary (such as the frontloading of US imports in response to the threat of high tariffs), some likely to be permanent (such as the decline in direct trade between the US and China) and some in between (such as the boom in AI-related trade). Nevertheless, global trade in goods, the products targeted by the tariffs, has been strikingly robust, according to Geopolitics and the geometry of global trade: 2026 update, a preliminary evaluation of 2025 by the McKinsey Global Institute.
The report notes five noteworthy facets of what happened in 2025.
First, US and Chinese exports reached new highs, while world trade also grew faster than the world economy. The direction of trade shifted substantially, but more from what McKinsey labels “geopolitically distant” trading partners — notably the US and China — than from “geographically distant” ones. Similarly, the EU lost market share in Chinese markets. But India stood out for a large increase in the geographical distance of trade because shipments of smartphones to the US rose fast. (See charts).
Second, AI-related shipments became the most powerful engine of trade in goods, with a rise of 40 per cent between 2024 and 2025 in the value of shipments of semiconductors and equipment for data centres. Indeed, AI-related exports accounted for one-third of global trade growth as Asian hubs — Taiwan, South Korea and parts of south-east Asia — supplied markets around the world, particularly the US. Controls on both exports and imports of some of this equipment restricted the growth of China’s AI-related trade to 16 per cent. The report argues that the rapid growth in AI-related capacity will continue to drive world trade in 2026.
Third, in the report’s words, “China expanded its role as a ‘factory to the factories’.” While its direct exports to the US came under attack, China was able to increase exports of machinery and inputs to other countries, notably its neighbours, some of which could then replace China’s exports to the US. In many other cases, notes McKinsey, its exports of parts and machinery were not tied to replacing lost sales to the US. Instead, they supported the expansion of manufacturing capacity in third markets, particularly emerging economies. This deepened China’s role as a supplier of inputs into production rather than as an exporter of final goods. In all, China’s exports of intermediate and capital goods rose by $223bn in 2025, more than offsetting a reduction of $130bn in exports to the US.
Fourth, the tariffs caused complex trade readjustments. Among other things, they caused temporary frontloading of imports. Overall, direct US-China trade fell by around 30 per cent in 2025. But the US replaced about two-thirds of the lost imports with purchases from other exporters, while Chinese exporters of consumer goods, such as electric cars and toys, cut prices by an average of 8 per cent to find new buyers. The Association of Southeast Asian Nations countries’ exports thrived in this new world. Meanwhile, says McKinsey, EU firms faced a double squeeze from diverted Chinese exports and higher US tariffs against their exports.
Finally, while the irrationality of President Donald Trump’s arbitrary and unpredictable tariffs made life difficult for producers and traders worldwide (inevitably including many in the US itself), there were a number of useful offsetting factors. One was that Trump’s bark was worse than his bite. In the end, as Richard Baldwin of IMD in Lausanne notes in his Substack post “Why Didn’t Trumpian Tariffs Wreck the World Trade System”, he did not do all that he threatened.
More importantly, his actions led neither to a cycle of retaliation against the US nor, crucially, to imitation of the aggressive US repudiation of World Trade Organization commitments and norms. The trading system has other big challenges, notably China’s aggressively mercantilist approach to exports. But the US, with just 14 per cent of world imports of goods, does not matter all that much. Indeed, even the US and China together, with 25 per cent of world trade between them, do not matter that much, as Baldwin notes in “How Did the 75-percenters Save the System?” on Substack. The rest of the world has decided to go on trading because it depends on it.
There is, it turns out, “a great deal of ruin” in world trade, as Adam Smith would have told us. But there may still be limits. Will the AI boom go bust this year? Might the impact of Trump’s war against Iran in 2026 surpass the damage done by his tariff war in 2025? More specifically, will the de facto closure of the Gulf to exports of oil, gas and other essential products end up causing more damage than we can manage? I suspect not. It seems likely that, however ill-conceived and ill-executed this war might be, Trump will find a way to claim victory and end it. Of course, it might be obvious to most that he has lost. But would that embarrass him? Probably not: he does not admit failure.
What is far clearer now, however, is that the US is no longer a credible world leader. If it can elect this man twice, it has lost the plot. Why would it not elect someone even worse? Such a country is incapable of providing reliable global hegemony. What it provides instead is an unpredictable wrecking ball. Nor is there a plausible replacement. China is relatively predictable. But its decades-long inability to eliminate reliance on huge trade and current account surpluses, to balance demand with excess domestic supply, is not encouraging.
What we have learnt is that the economy is more resilient than many feared. Let us hope that it continues to show that quality. It seems that we are going to need it, in spades.
