China’s economy grew faster than forecast in the third quarter of 2026, with official statistics showing GDP expansion of approximately 5.2 percent year-on-year — above the consensus analyst expectation of 4.6 percent and comfortably above the government’s official annual target. The BBC reported that the acceleration was driven primarily by a surge in AI-related exports — hardware, software platforms, and AI-embedded consumer electronics — combined with a recovery in domestic manufacturing that had been suppressed by the property sector downturn of the previous two years. The performance arrives as Washington and Brussels are in the middle of extended debates about the efficacy of semiconductor export controls imposed since 2022, and as the European Union separately weighs whether to impose tariffs on Chinese electric vehicles. It is a data point with implications that cut across trade policy, security strategy, and the increasingly contested question of whether the West’s technology containment policy has achieved its objectives.
The received wisdom
The standard liberal-internationalist framing treats China’s GDP beat as evidence of economic resilience that should inform a more measured engagement approach. The argument runs roughly as follows: export controls have not stopped Chinese AI development, they have merely imposed costs that Beijing has absorbed while accelerating domestic alternatives. Companies like Huawei and SMIC have developed chips that are less powerful than their American counterparts but functional enough for the current generation of AI workloads. The lesson, under this framing, is that confrontational decoupling is self-defeating — it costs Western companies market access, pushes China toward self-sufficiency faster than market competition would have done, and fails to achieve the security objectives that justified the controls. The better approach is targeted engagement: cooperate on AI safety governance, maintain competition on the frontier, and avoid the zero-sum logic that turns a trade relationship into a cold war.
A different read
The engagement argument is not frivolous, but it contains a category error that needs naming. The premise — that export controls have “failed” because China is still growing and still developing AI — conflates the metric of containment with the metric of competition. Export controls were never designed to stop China’s economy from growing. They were designed to slow the acquisition of the most advanced semiconductor manufacturing equipment and deny access to the highest-end chips used in frontier AI training. On that narrower metric, the evidence is more ambiguous than the GDP number suggests.
Analysts at the Center for Strategic and International Studies have noted that China’s AI export boom is concentrated in the mid-tier of the market — embedded AI in consumer electronics, logistics optimisation, facial recognition systems, and manufacturing robotics — rather than in the frontier large-language-model space where the most advanced American and British capabilities currently reside. The question is not whether China can build AI products that succeed commercially; it clearly can. The question is whether China can train the next generation of models at the frontier, which requires access to chips that export controls are still — imperfectly but meaningfully — restricting.
The European Union’s parallel calculation is worth examining separately. The BBC’s reporting noted that EU tariffs on Chinese goods are already affecting Ukraine — a steel manufacturer cited in that piece described how European protectionism designed to counter Chinese dumping was also raising costs for Ukrainian reconstruction. This is the uncomfortable geometry of trade policy: every tariff designed to protect a domestic industry or penalise a geopolitical rival creates collateral damage elsewhere in the supply chain. The EU’s EV tariff deliberations exist in this same geometry — protecting German and French automakers from Chinese competition while simultaneously making European consumers pay more for the clean-energy transition that climate policy demands.
The historical analogy that is most instructive here is not the Cold War semiconductor competition — which is the framing Washington prefers — but Japan in the 1980s. When Japan’s electronics and automotive industries appeared to be on a trajectory to dominate global markets, the US response combined protectionist pressure (the Plaza Accord, Voluntary Export Restraints) with financial liberalisation pressure that contributed to the asset bubble that ultimately devastated Japan’s lost decade. The combination of containment and engagement on American terms succeeded in slowing Japan’s challenge — but it also produced a twenty-year growth stagnation that impoverished ordinary Japanese citizens and created a geopolitical partner that has spent decades nursing resentments. China’s leadership has studied Japan’s experience extensively; the “do not repeat Japan’s mistake” lesson is part of Xi Jinping’s explicit economic doctrine. Whether that self-awareness means Beijing will navigate the current pressure more successfully, or whether it will simply delay the reckoning, is the central question.
The right-of-centre case for seriousness on this issue is not about hostility to China’s development per se. It is about institutional honesty: if export controls are not achieving their strategic objectives, say so and revise the policy. If they are buying time while Western AI labs maintain a capability lead, quantify how much time and use that time productively. The current policy sits in an uncomfortable middle position — too restrictive to allow the full commercial engagement that would benefit Western companies and expose China to the rule-of-law dependencies that liberal economists argue are stabilising, but not restrictive enough to actually prevent Chinese frontier AI development. That middle position may be the unavoidable result of coalition politics in Washington and Brussels, but it should at least be acknowledged as a strategic choice with costs.
What to watch
- The EU’s final decision on Chinese EV tariffs, expected before year-end: a high tariff will escalate trade tensions and invite Chinese countermeasures; a low tariff will test whether European industrial policy has the political staying power to shape the sector.
- Huawei’s next chip announcement: the company’s Kirin processor series has been the most visible indicator of how effectively export controls have contained Chinese semiconductor capability. A major leap forward would reset the policy debate.
- Whether the China GDP number changes any positions in the ongoing US Commerce Department review of export control implementation — which has been under internal criticism for creating paperwork burdens without strategic clarity.
- Domestic Chinese property sector data: the AI export boom is real, but the property debt overhang remains the most serious structural risk to sustained growth; if the sector deteriorates again, the Q3 number will look less durable.
— J