Everything About AI
NextFin News - China's benchmark CSI 300 index has entered a bear market after a 10-week rout erased roughly $3.5 trillion in stock-market value, but the deeper problem is not the price decline - it is the earnings slowdown underneath it. Falling corporate profits are turning what began as an artificial-intelligence trade unwind into a value trap that stimulus hopes alone are unlikely to fix.
The CSI 300 fell 1.8% on October 22, leaving it 20% below its June peak and marking the index's first bear market since the COVID crash of 2020. The sell-off has been the longest weekly losing streak in the index's history, and it is spreading across growth gauges: the Nasdaq-style Star50 index has tumbled about 40% in the second half of the year, with the ChiNext gauge down by a similar measure. Optical-module makers Eoptolink Technology and Zhongji Innolight were among the worst hit as the AI hardware complex unwound.
Yet the rout is not a simple sentiment story. As global enthusiasm for AI hardware fades and domestic demand stays weak, Chinese companies are posting slowing profit growth just as energy and input costs rise. Industrial profits grew only 4.2% in August, down from 11.2% in July and below the 8% forecast - the weakest pace since November 2025. When costs rise faster than the prices companies can charge, valuations that looked cheap on last year's earnings become expensive on next year's. That is the trap: investors buying the dip are catching a knife whose handle keeps getting shorter.
What the Profit Data Actually Shows
The headline industrial-profit number still looks respectable at first glance: profits at China's industrial firms rose 15.7% in the first eight months of 2026, to 5.27 trillion yuan. But that aggregate hides a sharp deceleration and a deep split inside the economy. August's 4.2% gain was the fourth straight month of slowdown from the 24.7% expansion recorded in April, and the weakest monthly reading since November 2025.
The bifurcation is the point. Mining profits surged 35.1% and computers and communications equipment jumped 110%, powered by elevated energy prices and the global AI build-out. Meanwhile consumer-facing industries - clothing, autos, furniture - saw profits decline. Utilities fell 12%. In other words, the winners are upstream suppliers and export-linked tech hardware; the losers are the companies that depend on Chinese households opening their wallets.
That pattern is a margin squeeze, not a demand boom. Producer prices rose 5.0% year-on-year in September, up from 3.8% in August, as higher fuel costs fed through the supply chain. Manufacturers facing weak consumer demand cannot fully pass those costs on, so the difference comes out of gross margin. The official data showed industrial value added grew 5.3% in the first three quarters - output is holding up even as profitability thins. This is the classic profit-squeeze dynamic: revenue keeps growing, but each additional yuan of sales earns less than the last.
How a Margin Squeeze Becomes a Valuation Trap
The transmission from profit slowdown to equity rout runs through the denominator of the price-to-earnings ratio. Investors see a stock trading at a low multiple and call it cheap. But the multiple is priced against trailing earnings. If profits fall, the trailing multiple mechanically rises even if the share price does nothing - and if the market then de-rates the stock to reflect weaker future earnings, price and multiple fall together. That double hit is what turns a "cheap" market into a value trap.
The CSI 300 trades at about 12.6 times trailing earnings, below its long-term median of roughly 13.9 and far below the 20.4 peak reached in June 2015. On the surface that looks like a steep discount. But the index has already erased about $2.4 trillion in market value since its May peak, and some $6.3 trillion since the 2021 high, without earnings growth arriving to justify a re-rating. A market can stay cheap for years when the earnings engine is stalling. The valuation is not a floor; it is a symptom.
The wealth channel reinforces the squeeze. Chinese households hold the bulk of their net worth in real estate, and the property correction that began in 2021 has destroyed the main store of household wealth. When the apartment stops appreciating, families spend less and demand more from their savings. That is why retail sales rose just 1.1% in the first nine months of the year despite a 4.6% GDP expansion. Weak consumption is not a cyclical blip here; it is the arithmetic of a balance-sheet recession playing out in slow motion.
Why the AI Rally Broke First
China's tech hardware complex was the first to crack because it had the furthest to fall. Early optimism about domestic AI breakthroughs and state support drove valuations to lofty levels, leaving the sector exposed when global investors turned skeptical about AI spending payoffs. Mixed results from U.S. megacaps intensified the scrutiny: if the American leaders cannot monetize AI fast enough, then the Chinese suppliers selling them optical modules and servers face an order slowdown before their own earnings catch up.
The trigger was CATL. The world's largest EV battery maker reported third-quarter net profit of 21.25 billion yuan, up 14.5% year-on-year - well below the 41.2% expansion analysts had expected, and the slowest quarterly profit growth in nearly two years. Revenue rose 58.6% to 165.22 billion yuan, but gross margin fell to 19.3%, pressured by product-mix changes and a reinstated battery consumption tax. CATL's shares dropped 12%.
"Investor sentiment remains very poor for China's onshore and offshore markets, and without new policy or tech catalysts, this kind of volatile, directionless environment may persist into early 2027," said Homin Lee, a senior macro strategist at Lombard Odier in Singapore.
The CATL print mattered because it converted an abstract worry - "AI profitability is uncertain" - into a concrete data point: margin compression at the best-capitalized company in the supply chain. When the strongest player slows, the market reprices the whole chain.
The Stimulus Disappointment
The second leg of the rout is policy. After Beijing's late-2024 stimulus push drew investors back to Chinese equities, traders spent 2026 expecting a repeat: a large, 2024-style fiscal package to lift domestic demand. Policymakers instead stuck with targeted and incremental measures - enough to stabilize growth, not enough to reflate earnings.
The economy is not collapsing, which is precisely why Beijing is not panicking. Third-quarter GDP grew 4.5% year-on-year, and the first nine months came in at 4.6%, on track to meet the official 4.5%-5% target. Exports surged more than 16% in the first three quarters, driven by semiconductors and other tech goods. Retail sales, the domestic-demand proxy, rose just 1.1% over the same period.
That mix - strong external demand, weak household spending - is why incremental stimulus feels insufficient to equity investors. A factory making AI servers for export does not need Beijing's help; a clothing manufacturer watching domestic orders shrink does. The policy that has arrived helps the former; the policy the market wanted would have targeted the latter.
State-backed buying has stepped in to blunt the fall, but it is a floor, not a recovery. UBS calculated that state players bought a net 410 billion yuan, or $57 billion, of exchange-traded funds this year, with 76% of that funnelled into ETFs tracking the CSI 300. Further state-fund intervention was reported on October 23 after shares slid into bear territory. The buying supports the index, but it does not create earnings. A market held up by official bids is stable only for as long as the bids continue.
Cyclical Slowdown or Structural Shift?
The central question is whether this is a cyclical earnings dip that will revert, or a structural change in how Chinese equities compound value. The evidence points to both operating at once - and that is what makes the trap so dangerous.
The cyclical leg is real and, by historical precedent, mean-reverting. China's industrial profits have swung through sharp down-cycles before and snapped back: profits fell through 2019 as the trade war bit, declined for a third straight year in 2024 with a 4.7% drop in the January-November period, and fell 2.8% in the first half of 2025 before rebounding into double-digit growth this year. The pattern is classic late-cycle: input costs rise, pricing power fades, margins compress, inventories clear, and the cycle turns. If global growth holds and domestic stimulus eventually arrives in larger size, earnings growth can recover within a few quarters.
But the structural leg is different, and it does not self-correct. Chinese households' share of national income has been declining for years, and the property correction has destroyed the main store of household wealth. When the asset that families rely on - the apartment - stops appreciating, consumption stays cautious no matter how many targeted subsidies arrive. That is a regime change in the growth model, not a cycle. It means the consumer-facing sectors that led the 2019 and 2025 recoveries may not pull the next one with the same force.
The bifurcation in the profit data is the structural signal. The economy is splitting into an externally oriented, state-supported tech-and-export complex that grows, and a domestic-consumption complex that stagnates. An index that owns both will not rebound as a unit. Picking the "cheap" index is no longer a bet on China's growth; it is a bet on which half of the economy wins.
The Counter-Thesis: This Is Just a Sentiment Washout
The strongest case against the trap thesis is straightforward: Chinese equities are already priced for disaster, and the bad news is fully reflected. The CSI 300 trades at about 12.6 times trailing earnings, below its long-term median of roughly 13.9 and far below the 20.4 peak reached in June 2015 - a steep discount to U.S. and Japanese peers. State-backed buying has already absorbed $57 billion of selling pressure this year, and Beijing has a deep toolkit - rate cuts, reserve-requirement reductions, direct equity purchases - that it has not fully deployed. If policymakers deliver a meaningful fiscal package at the next major meeting, the rebound could be violent precisely because positioning is so bearish.
This is a serious argument, and it underpins any contrarian buy. But it rests on two assumptions that the profit data undermines. First, it assumes cheap valuations are a floor. They are not, when earnings are falling: a stock at 12 times earnings that sees profits drop 20% is at 15 times on the new run-rate, and the multiple can keep expanding even as the price keeps falling. Second, it assumes policy will act decisively. The authorities have shown a clear preference for targeted support while growth remains near target - and with GDP on pace for the official goal, the pressure to unleash a 2024-style bazooka is limited.
The falsifying signal for the trap thesis is specific: if industrial profits re-accelerate to double-digit year-on-year growth for two consecutive months while retail sales growth picks up above 4%, the margin-squeeze narrative is wrong and this is a cyclical dip offering a buying opportunity. Until then, the burden of proof sits with the bulls.
What Comes Next
In the short term, the market is a function of liquidity and sentiment, not fundamentals. With sentiment poor and no policy catalyst on the immediate calendar, range-bound volatility into early 2027 is the base case. A state-fund intervention or a surprise stimulus announcement could produce a sharp relief rally - tradable, but not necessarily durable, because the state buys indexes, not confidence.
Over the medium term, earnings are the tiebreaker. The January-September profit data, due later this month, will show whether August's 4.2% slowdown was a one-off or the start of a deeper trend. Investors should watch three numbers: the nine-month industrial profit print, the gross-margin trajectory at export champions like CATL, and retail-sales growth. A second consecutive month of single-digit profit growth would confirm the squeeze is broadening.
In the long term, the structural question dominates: can China rebalance toward household consumption fast enough to give its equity market a broad earnings engine, or does it remain dependent on exports and state-directed investment? The answer determines whether today's lows are a generational buying opportunity or a value trap that grinds lower in real terms.
Base case: no big stimulus, earnings growth stays in the low single digits, and the CSI 300 remains directionless with a downward bias. Upside case: a meaningful fiscal package plus a global AI re-acceleration sends the index back toward its June peak. Downside case: profit growth turns negative and the bear market deepens toward 30% below the peak.
The trap is not that China's stocks fell. It is that the fall looks like a discount when the earnings underneath it are still slowing - and a cheap stock with falling profits is the most expensive position in the market.
Explore more exclusive insights at nextfin.ai.
