China's economy grew 4.5% year-on-year in Q3 2026, beating market expectations and accelerating from Q2's 4.3%. The National Bureau of Statistics figures put cumulative 2026 growth at 4.6% through nine months, squarely within the government's 4.5%-5% target range. Beijing will read this as vindication — no additional stimulus required. The engine is exports, specifically AI-related equipment and electronics. September exports surged more than a third year-on-year, with manufacturing output growing 6.0% and hi-tech manufacturing outperforming traditional sectors. Duncan Wrigley of Pantheon Macroeconomics described China as "climbing the value-added ladder," a polite way of saying the country's industrial policy is working precisely as designed — moving up the tech stack while the rest of the world buys the output. The domestic side tells a different story. Retail sales grew just 1.1% in September, up from a dismal 0.4% in August. Fixed-asset investment fell 7.2% over the first nine months. Property investment cratered 20.6%. Infrastructure spending remained weak. This is not an economy where households are spending freely — it is an economy where factories ship goods abroad while consumers sit on their wallets. The structural asymmetry is striking. Industrial production rose 5.5% in September while retail barely moved. The gap between what China makes and what Chinese people buy continues to widen. ABN Amro analysts attributed the GDP beat to "ongoing export strength on the global AI/tech boom, more than offsetting drags from the Iran conflict and ongoing weakness in domestic demand." The offsetting is doing heavy lifting in that sentence. Beijing's calculus is clear: as long as headline GDP hits target, the consumption problem can wait. The government's industrial policy channels credit, subsidies, and regulatory support toward manufacturing and exports. Households absorb the costs through stagnant wages relative to output, a weak property market that has destroyed household wealth, and an implicit subsidy to producers over consumers. The global AI boom provides convenient cover. When DeepSeek and its peers drive demand for Chinese-made chips, servers, and electronics, policymakers can point to a growth number that looks healthy without addressing the demand deficit underneath. But export dependency at this scale creates fragility — tariffs, geopolitical disruption, or a slowdown in global AI capital expenditure would expose the single-engine problem immediately. The twenty-year version of this pattern is a manufacturing superpower with a suppressed consumer class, increasingly vulnerable to external demand shocks and trade barriers. China has run this playbook before. The question is whether the AI export surge is a bridge to rebalancing or an excuse to avoid it.
