NextFin News - China's listed companies just delivered their strongest earnings in years, and the stock market barely blinked. After a first half of 2026 in which the country's roughly 5,557 reporting companies booked a combined 3.58 trillion yuan in net profit, up 19.5 percent from a year earlier, the CSI 300 index of mainland blue chips closed September 4 at 4,548.05, down 2.36 percent over the prior four weeks and up less than 2 percent across twelve months. The disconnect is the story: corporate profits are roaring back, but investors are treating the rebound as a cyclical bounce, not a reason to re-rate China risk. The market is asking a harder question than "are earnings good?" - it is asking whether this earnings strength can survive without a recovery in domestic demand, and whether the geopolitical and property headwinds that crushed valuations have actually gone away.
The Numbers Behind the Boom
The earnings tally, released September 2 by the China Association for Public Companies, is the clearest hard evidence yet that the world's second-largest economy is stabilizing at the corporate level. Net profit of 3.58 trillion yuan, or about 528 billion U.S. dollars, marks the first time the half-year figure has crossed the 3.5 trillion yuan threshold. About three-quarters of the 5,557 companies that published interim reports remained profitable, and growth accelerated through the period, with second-quarter revenue and net profit both rising markedly faster than in the first quarter.
The headline number, however, understates the force of the rebound in the sectors that matter to investors. Growth is heavily concentrated in the new-economy complex. Companies listed on the STAR Market, home to China's semiconductor and hard-tech champions, saw net profits surge 4.4-fold year on year on revenue growth of nearly 40 percent. The ChiNext board of Shenzhen-listed growth companies reported a 32.7 percent increase in net profit, with revenue up 22.3 percent. Private companies, the most sensitive gauge of domestic business confidence, posted net profit growth of 29.6 percent.
Integrated circuits led the charge with net profits up 2.4-fold, as the artificial-intelligence computing boom fed demand for memory, optical modules and AI chips. High-end equipment manufacturing reported revenue growth of 13.1 percent and net profit growth of 16.7 percent. Ten of the 19 major industry categories recorded simultaneous growth in both revenue and profit, a breadth that matters because it suggests the recovery is not being carried by a handful of commodity names.
Individual standouts read like a roll call of the AI supply chain. CXMT, the memory chipmaker that listed on the STAR Market in July, swung from a loss of 2.3 billion yuan a year earlier to an H1 net profit of 77.6 billion yuan, with revenue up 873.6 percent. Foxconn Industrial Internet, the AI server manufacturer, saw net profit rise 96 percent. Zhongji Innolight, the optical-module supplier, posted net profit of 13.65 billion yuan, up 241.7 percent, while AI chip designer Cambricon saw net profit surge 122.6 percent. Beyond the chip complex, battery maker CATL reported H1 net profit of 43.3 billion yuan, up 42 percent.
The quality of the earnings also improved on paper. Research and development spending by listed companies totaled 847.3 billion yuan in the first half, up 3 percent year on year, with overall R&D intensity at 2.24 percent; on the STAR Market, R&D intensity has exceeded 10 percent of revenue for several years. Shareholder returns stepped up as well: 872 companies announced interim cash dividends totaling 740.3 billion yuan, and 1,051 companies planned share buybacks worth more than 220 billion yuan.
Against this backdrop, the equity indices look almost sullen. The CSI 300 was down 0.10 percent on September 4, the Shanghai Composite at 3,930.12 was down 0.30 percent, and Hong Kong's Hang Seng, despite a 1.43 percent bounce that day, had still lost 1.32 percent over the prior month. A market that believes earnings growth is durable does not trade at a discount to its own history. The gap between the profit report and the price action is where the real story sits.
Why the Market Is Not Buying the Boom
The first reason the rally did not materialize is that the earnings rebound is being read as cyclical, not structural - and cyclical rebounds in China have a habit of rolling over before they reach the broader economy. The single most important driver of the profit surge is the exit from factory-gate deflation. China's Producer Price Index, after 41 months of year-on-year decline, has finally turned positive. That mechanically lifts reported profits for manufacturers and commodity producers without requiring any increase in real sales volume. When prices stop falling, margins recover even if demand stays flat.
This is the mechanism investors are pricing, and it is a mean-reverting one. A deflation exit is a one-time level shift in the price path, not a permanent step-up in the growth rate. Once the base effect rolls through, the contribution to profit growth fades. That is why the same data release that showed 19.5 percent earnings growth also showed industrial profit growth cooling to 11.2 percent year on year in July, down from 15.1 percent in June and 18.7 percent for the first half. The peak rate of improvement may already be behind the market.
The second reason is transmission. Corporate profits have improved, but the benefits have not clearly reached households, and household spending is what China's economy needs to rebalance. The overall economy expanded 4.7 percent in the first half, and overseas revenue of 3,196 listed companies rose 22.9 percent to 6.06 trillion yuan - but that is an external demand story, not a domestic one. With 553 companies generating more than half their revenue overseas, a large share of the profit boom is tied to global AI capital expenditure and to markets outside China. Investors who worry that tariff escalations could reverse that export engine are not going to pay up for earnings that depend on foreign buyers.
The third reason is valuation risk, not valuation comfort. Chinese equities are cheap on a price-to-earnings basis, and bulls cite that as the reason to buy. But a low multiple is only a catalyst if the earnings denominator proves durable. If earnings are cyclical and peak in 2026, today's "cheap" multiple is tomorrow's value trap. The market is effectively saying: show me that this profit growth survives a tariff shock and a property drag, then we will re-rate.
The Bull Case: This Time, the Earnings Are Real
The counter-thesis is not weak, and it deserves to be stated at full strength. The bears are right that the PPI turn is a base effect - but they are wrong to dismiss what sits underneath it. This is not a broad-based reflation driven by stimulus; it is a concentrated profit surge in sectors where China is genuinely gaining global share. Integrated circuits up 2.4-fold, optical modules up 241 percent, AI servers up 96 percent - these are not accounting artifacts. They are the financial statements of an industrial upgrade that is actually happening.
Li Xunlei, chief economist at Zhongtai Financial International Limited, framed the pattern as a concentrated surge with broad participation, attributing the profit growth across semiconductors, communications and electronics to the AI computing boom. He stressed that the AI boom is "by no means a short-term bubble but a confirmation of a long-term industrial upgrade trend," and forecast a 12 to 15 percent increase in non-financial corporate earnings for the full year.
Ruan Xiaoqin, head of the policy research office at SWS Research, pointed to deeper changes beneath the headline growth. A-share listed companies saw gains in both revenue and net profits, presenting what she described as a sound picture of overall recovery, structural improvement and renewed growth momentum:
Structural improvement was another keyword, with profit, industrial and market value structures all upgraded.
She added that non-financial manufacturing and hard-tech growth industries have become the main contributors to profit growth, while large-cap companies on the main board remain broadly stable.
There is also the matter of positioning. Global equity funds remain underweight China relative to recent averages, and domestic households hold an estimated 163 trillion yuan in savings that strategists say have been parked in bank deposits as property prices fell. If even a small fraction of that savings pool rotates into equities in search of yield - as deposit rates decline and the property route to wealth creation stays broken - the flow of funds alone could overwhelm the cautious earnings narrative. Goldman Sachs, which called the 2026 rally early, has raised its CSI 300 target to 5,300 points and expects 20 percent earnings growth for A-shares in 2026, with market consensus for 2026 earnings growth revised up from 16 percent to 23 percent.
The bulls' strongest point is this: markets usually discount the future, and if the future is a China where hard-tech earnings compound at double-digit rates while the rest of the economy muddles through, today's prices are discounting the muddle and ignoring the compounding. The selective rally in AI hardware names is not irrational - it is the market pricing exactly that bifurcation.
What Would Prove Each Side Wrong
The bear case rests on one falsifiable proposition: that the earnings rebound is cyclical and will fade as the PPI base effect rolls off. The specific signal that would break the bear case is full-year 2026 non-financial corporate earnings growth coming in at or above the 12 to 15 percent range forecast by Zhongtai's Li Xunlei - not from price recovery, but from volume and market-share gains that persist into 2027. If earnings growth holds above 12 percent after the base effect has fully passed, the "cyclical bounce" label no longer fits.
The bull case rests on the opposite proposition: that this earnings strength is durable enough to justify a valuation re-rating. The signal that would break the bull case is simpler and closer at hand. If the CSI 300 trades below roughly 4,400 while earnings reports continue to beat expectations - meaning investors are rewarding good news with selling - then the market has made its judgment: the earnings are real, but China's risk premium is not coming down. A second falsifying signal for the bulls is full-year earnings growth below 12 percent, which would confirm that the PPI base effect did most of the work.
What to Watch Next
The near-term path is likely to be range-bound and selective. The CSI 300 has been oscillating in a zone between roughly 4,500 and 4,700 through the recent sessions, and there is no obvious catalyst to break out of it without a policy signal on domestic demand or a de-escalation in trade tensions. In that environment, the market will continue to reward the AI supply chain and hard-tech exporters while discounting domestically exposed cyclicals and property-linked names.
Three data points will decide the next leg. First, the full-year 2026 earnings growth print for non-financial corporates - the 12 to 15 percent forecast range is the line between "cyclical bounce" and "sustained recovery." Second, the PPI and CPI path through the fourth quarter - if factory-gate inflation stays positive and consumer prices firm, the margin recovery has a second leg; if PPI rolls back toward zero, the base effect is exhausted. Third, policy: any concrete fiscal measure aimed at household income or consumption, rather than infrastructure and supply-side support, would be the clearest signal that Beijing is closing the transmission gap between corporate profits and domestic demand.
Scenarios are easy to state and hard to bet on. The base case is a range-bound market where earnings provide a floor but not a ceiling - selective upside for AI hardware, semiconductors and exporters, stagnation for the rest. The upside case requires the confluence of domestic stimulus, property stabilization and tariff de-escalation; if all three arrive, the Goldman 5,300 target on the CSI 300 becomes reachable. The downside case is a tariff escalation that hits the 22.9 percent export-revenue growth engine while the property drag persists - in that world, earnings revisions turn negative and the CSI 300 tests the 4,000 to 4,200 zone.
The split by time horizon is the cleanest way to frame the call. Over one to three months, sentiment and liquidity dominate: expect volatility, not trend. Over six to twelve months, fundamentals provide a floor - earnings this strong do not collapse without an external shock, and that supports selective longs. Over one to three years, the structural question dominates: Chinese equities will not sustainably re-rate until domestic demand recovers, the property overhang clears, and the geopolitical risk premium compresses. None of those three has happened yet.
The market is not saying China's earnings are bad. It is saying that good earnings, by themselves, are not enough to make China a buy. Until profit growth proves it can travel from the factory gate to the household wallet, the boom will keep landing with a thud.
Explore more exclusive insights at nextfin.ai.
