NextFin News - China's economy slowed to its weakest quarterly pace in more than three years in the second quarter, yet the country's factories are running hotter than they have in years - powered by a single force the world cannot get enough of: artificial intelligence. The divergence is the defining story of 2026 for the world's second-largest economy. Domestic demand is too weak to carry growth, so China is leaning on an AI-fueled export boom that is reshaping what it sells, who it sells to, and how exposed it is to the politics of its trading partners.
Data released by the National Bureau of Statistics on September 15 showed the domestic engine still sputtering. Retail sales rose just 3.4% in August from a year earlier, missing the 3.9% growth economists expected and slowing from July's 3.7%. Industrial output growth cooled to 5.2%, the weakest reading since August 2024, while fixed-asset investment for the first eight months of the year expanded a mere 0.5%. Real estate investment, the old growth machine, contracted 12.9% over the same stretch. Yet in the very same week, customs data painted a different country: exports jumped 23.9% in July to $397.85 billion, and the trade surplus widened to $112.5 billion, well above the $107 billion forecast. The question is not whether the export boom is real. It is whether an economy the size of China's can outsource its recovery to the rest of the world's AI spending - and at what cost.
The Two Chinas: A Weak Home Market, a Red-Hot Factory Gate
The numbers tell the story of an economy split in two. On the consumption side, Chinese households are pulling back. August retail sales excluding automobiles grew 3.7%, and the consumer price index fell 0.4% year-on-year - a deflationary signal that households are still unwilling to spend freely even as the government rolls out trade-in subsidies for cars and home appliances. Producer prices have now been in deflation for a third consecutive year, evidence that factories are competing ferociously on price rather than passing costs through.
The property crisis remains the anchor dragging on the domestic side. Real estate has historically been the single largest store of household wealth in China, and a multi-year downturn in housing values has rippled outward into spending decisions and local government revenue. Fixed-asset investment is being held up almost entirely by state-owned enterprises; private-sector investment contracted from a year earlier, according to Yuhan Zhang, principal economist at The Conference Board's China Center. When the private sector - the part of the economy that creates jobs and takes risks - is retreating while the state steps in, growth becomes an accounting exercise rather than a self-sustaining cycle.
Contrast that with the external side. In June, exports surged 27% year-on-year, the strongest performance in four months and far above the 18.2% economists had forecast. Imports jumped 36%, a five-year high. The trade surplus for June alone reached $125.6 billion, up from $105.4 billion in May. For the first half of 2026, total exports rose 17.6% to $2.1253 trillion, with high-tech exports - the fastest-growing major category - up 39% to 3.26 trillion yuan. July's data confirmed the trend was not a one-off: high-tech exports rose 40.7% and semiconductor shipments nearly doubled in value terms compared with a year earlier.
This is not a broad-based export recovery. It is a narrow, concentrated boom in the products that feed the global AI buildout - chips, data-processing equipment, and the electronics that house them. The mechanism is straightforward: every data center being built in the United States, Europe, and Asia needs computing hardware, and China's manufacturing base sits at the middle of that supply chain even as it faces tariffs and technology restrictions. The irony is sharp. The same geopolitical rivalry that produced US export controls on advanced chips has made China's role as the workshop for AI infrastructure more valuable, not less - at least in the near term.
"Continued export strength, mostly driven by AI, points to a better second half, coupled with a more expansionary policy mix, accelerated fiscal spending and mild monetary easing," said Xu Tianchen, a senior economist at the Economist Intelligence Unit in Beijing.
But Xu's own assessment captures the asymmetry: "domestic demand remains a drag. Retail sales remain pretty flat and fixed-asset investment was negative last month." One side of China's economy is being carried by a global technology supercycle. The other is sitting still.
Why the Export Boom Will Not Fix What Is Broken at Home
The first-order reading of the data is comforting: exports are strong, so growth will hold. The second-order question is whether export strength can substitute for domestic demand in an economy of China's scale - and the answer points to a structural limit, not a cyclical gap.
The transmission channel runs through dependence. The ratio of annual exports to total manufacturing sales hit 24% in the first four months of 2026, according to consultancy Gavekal Dragonomics - the highest level since China joined the World Trade Organization in 2001. In 2019, that ratio stood at 18.3%; it rose to 22.3% in 2025. For a small, export-focused economy, 24% would be unremarkable. For the world's second-largest economy, with 1.4 billion people and an official growth target of 4.5% to 5% for the year, it is a sign that the rebalancing Beijing promised more than a decade ago - away from exports and investment, toward consumption - has gone into reverse.
Here is the structural call this piece defends: China's domestic weakness is structural, not cyclical. A cyclical slowdown mean-reverts when policy loosens or inventories clear. What China faces is a regime change in its growth model. The property sector, which at its peak accounted for roughly a quarter of economic activity when construction and related industries are counted, is not coming back. Household balance sheets are tied to falling home values. Local governments, dependent on land sales, have less fiscal room. Deflation has taken hold, and deflation raises the real burden of debt - which is why the statistics bureau's own spokesperson, Fu Linghui, acknowledged that "national economic development is still confronted with multiple risks and challenges" while stopping short of promising a large stimulus.
The export boom, by contrast, is cyclical riding on a structural reorientation. The structural part is real and durable: China has successfully repositioned its manufacturing base toward higher-value technology goods, and its dominance in the middle of the AI hardware supply chain is not easily displaced. But the current intensity of the boom is being amplified by two cyclical forces that will fade. First, the global AI capital-expenditure supercycle - a spending wave that, like every capex cycle, will eventually slow as data-center capacity catches up with demand. Second, tariff front-running: the United States applied a new 12.5% levy on Chinese products in late July, replacing a temporary 10% levy that had lapsed, and importers have been pulling shipments forward to beat higher duties. Pull-forward demand borrows orders from the future.
Julian Evans-Pritchard, head of China economics at Capital Economics, warned against reading too much into the import surge that accompanied the export boom. "As with exports, surging semiconductor prices are playing a key role in pushing up import values," he said - meaning part of the trade data reflects price inflation in chips, not a broad recovery in Chinese demand. Imports from South Korea rose 85% in June and those from Taiwan rose 41.1%, both major semiconductor suppliers. That is the AI supply chain humming, not the Chinese consumer returning.
The consequence is a policy trap. If exports are doing the heavy lifting, Beijing has less incentive to unleash the kind of household-transfer stimulus that would actually rebalance the economy toward consumption. And if Beijing does not rebalance, the export dependence deepens - which in turn deepens the political backlash abroad. Zhiwei Zhang, chief economist at Pinpoint Asset Management, put it plainly: "I think exports will remain strong in the second half of the year. Meanwhile, it also puts further pressure on the trade tensions between China and its trading partners, Europe in particular."
Short-knife close: the export boom is not a cure for China's domestic weakness. It is a painkiller that makes the underlying condition easier to ignore - and harder to treat.
The Counter-Thesis: Why the Boom Could Last Longer Than Skeptics Expect
The strongest case against the view above is that it underestimates how long and deep the AI investment cycle really is. This is not a normal capex upswing. Artificial intelligence is a general-purpose technology, and the infrastructure buildout required to deploy it - data centers, power grids, networking equipment, advanced packaging - could take a decade or more to complete. If global AI spending proves to be a structural shift rather than a cyclical wave, then China's position at the center of the hardware supply chain is a structural advantage, not a temporary tailwind.
There is also evidence that China's export machine is diversifying its destinations, reducing its vulnerability to any single market. Shipments to the United States rose 17% year-on-year in July, customs data showed, even as the two countries remain locked in a trade war - suggesting that Chinese exporters are routing through third countries and that US demand for Chinese-made AI inputs is more inelastic than policymakers assume. Meanwhile, the Belt and Road economies and the Global South offer alternative markets that are less politically sensitive to Chinese overcapacity.
Finally, the counter-thesis notes that Beijing still has substantial fiscal and monetary ammunition. The central government's balance sheet is comparatively healthy, and special sovereign bonds, policy-bank lending, and targeted rate cuts can keep infrastructure and high-tech investment growing even if the property sector and private consumption do not recover. In this reading, China does not need a consumer-led rebalancing to hit its 5% target; it needs the state to keep spending and the world to keep buying chips.
This case has force, and it is the base case for a growing number of institutions - the International Monetary Fund, for example, recently raised its 2026 growth forecast for China by 0.2 percentage point to 4.6%. But it rests on two assumptions that deserve scrutiny. First, it assumes the rest of the world will tolerate a Chinese trade surplus on track to exceed $1 trillion for a second straight year. Tariffs are the polite version of the backlash; anti-dumping duties, local-content rules, and technology blacklists are the impolite versions, and all are already in motion. Second, it assumes that AI hardware margins will remain attractive as competition intensifies. China's own "anti-involution" campaign - Beijing's effort to curb the price wars and overcapacity that have gutted profits in sectors from solar panels to electric vehicles - is an admission that Chinese industry's instinct to cut prices and grab share can destroy the very export profitability the economy now depends on.
The falsifying signal for the view presented here is specific and observable: if China's exports continue to grow at double-digit rates through the first half of 2027 and the exports-to-manufacturing-sales ratio climbs above 26% without triggering significant new trade barriers from the European Union or the United States, then the AI export boom is durable enough to carry China's rebalancing burden, and the structural-pessimism case is wrong. If instead tariffs escalate, the AI capex cycle rolls over, or the ratio stabilizes while domestic consumption stays flat, the current path is not sustainable.
What Comes Next: Beneficiaries, Risks, and the Signals to Watch
Translating the mechanism into concrete implications, the picture splits by time horizon.
In the short term - the next two to three quarters - the export boom provides a floor under growth. Second-quarter GDP came in at 4.3%, the weakest quarterly pace since the fourth quarter of 2022 and below the 4.5% consensus, bringing first-half growth to 4.7%. With exports running at this pace, China is likely to clear the low end of its 4.5% to 5% annual target range without dramatic new stimulus. Beneficiaries are clear: Chinese technology exporters, semiconductor supply-chain firms, and the ports and logistics companies that move the goods. Global AI companies benefit too, through cheaper and more available hardware inputs - which is one reason the AI boom and China's export numbers are two sides of the same coin.
In the medium term - one to two years - the risk shifts to the external reaction function. A trade surplus of this size, concentrated in the sectors that Western governments consider strategically sensitive, is a political target. Europe is the most exposed pressure point: its own manufacturers are struggling against Chinese competition in electric vehicles and green technology, and Chinese tech exports give Brussels more reason to act. The exposed parties are Chinese exporters without diversified markets, and any global investor long Chinese equities on the assumption that exports will keep rising in a straight line.
In the long term - three years and beyond - the structural question dominates. China either rebalances toward household consumption, which requires transferring income to households and accepting slower headline growth in exchange for more sustainable growth, or it remains dependent on external demand and accepts the political friction that comes with it. There is no middle path where China runs a $1 trillion-plus trade surplus indefinitely while its trading partners stay quiet.
Three scenarios frame the outlook. The base case: exports moderate to high-single-digit growth as the AI capex cycle normalizes and tariff front-running fades, GDP grinds out 4.5% to 4.7% for the year, and trade tensions simmer without exploding. The upside case: the AI supercycle extends well into the late 2020s, China successfully diversifies exports into emerging markets, and Beijing engineers a soft landing at around 5% with targeted stimulus. The downside case: a sharp escalation of tariffs from the US and EU, combined with a global AI spending slowdown, pushes export growth into the low single digits while domestic demand remains flat - in which case the 5% target becomes unreachable without stimulus on a scale Beijing has so far refused to contemplate.
The signals to watch are concrete. On the domestic side: whether retail sales growth re-accelerates above 4% and whether the CPI returns to positive territory, both signs that the consumption turn has begun. On the external side: the monthly export growth rate and the direction of the exports-to-manufacturing-sales ratio; any European Union announcement of new restrictions on Chinese technology imports; and the US tariff schedule beyond the current 12.5% levy. And on policy: whether Beijing deploys direct household support - the clearest signal that it is serious about rebalancing - or continues to rely on state-led investment.
China's AI-fueled export boom is real, powerful, and buying the economy time. But time is not the same thing as a solution. The surplus that is keeping growth afloat today is also the imbalance that will define China's economic diplomacy for the rest of the decade - and the bill for depending on the world's AI spending to solve a domestic demand problem will eventually come due.
Explore more exclusive insights at nextfin.ai.
