Chinese onshore-listed companies reported a 25.7% rise in profits for the three months ended June, the fastest pace in nearly five years. The headline number, however, failed to reassure equity investors. The CSI 300 Index has fallen about 9% during the current quarter, while the tech-heavy Star 50 Index has dropped roughly 29%. Both benchmarks had rallied strongly in the previous period; the Star 50 had jumped 76% before the reversal.
A narrow earnings boom
The aggregate profit increase was not spread evenly across the market. According to an estimate from a leading brokerage, earnings growth reached about 42% on the ChiNext growth enterprise board and more than 370% on the STAR Market, the technology-focused board. The main board trailed by a wide margin. This distribution clarifies the market’s response: most of the improvement came from companies that are closely tied to artificial intelligence and related infrastructure, leaving a broader industrial recovery still missing.
A narrow rally in earnings, combined with an already crowded trade, made the technology sector especially vulnerable. After a spectacular run, much of the optimism was already embedded in prices. When companies finally delivered the strong numbers investors had waited for, the response was to sell, not to buy.
Strong numbers no longer work for tech
The old relationship between better earnings and higher share prices has broken down in China’s technology space. A managing director at a Swiss private bank summed up the mood by saying that strong numbers no longer work for technology stocks. The reasons include uncertainty over how much companies will need to invest in AI, the absence of a clear return on investment, and rising financing costs.
Heavy capital expenditure is now showing up in reported profits. Alibaba, for instance, traded lower in Hong Kong after issuing results that showed higher revenue but sharply lower profit. The company attributed the pressure to the cost of AI projects and computing infrastructure. It is raising $10.2 billion to continue spending in the same areas. In other words, the market punished a company for funding a future that investors themselves have demanded.
Tencent moved in a similar direction after more than doubling its AI-related spending. Its capital expenditure rose by 176% to 52.8 billion yuan in a recent period, and free cash flow turned negative by 13.8 billion yuan. On their own terms, the results were not catastrophic, but the market’s patience with heavy upfront spending appears to have thinned.
Domestic demand and currency drags
Company earnings were also shaped by domestic conditions. Consumer demand remains weak, and the property sector is still in a prolonged downturn. A stronger yuan added another complication. Exchange losses at non-financial A-share companies reached 107 billion yuan in the first half of the year, reducing profits for exporters and for firms with large foreign-currency liabilities.
The broader economy has been recovering unevenly. Government stimulus measures have supported industrial production and exports, but confidence among households and small businesses has been slower to return. This divergence explains why many businesses outside the AI supply chain have yet to report the kind of earnings growth that their share prices once anticipated.
A liquidity overhang
Another factor weighing on Chinese equities is the pipeline of new listings. The sale of new shares drains liquidity from the secondary market, particularly for companies that have already appreciated sharply. Yangtze Memory, a major memory chip maker, is among the large offerings still expected. Tax enforcement has also tightened, adding pressure on companies that previously had more room to manage their reported profits.
New share issuance is not new, but its timing matters. When the market is already nervous about AI valuations, a fresh supply of shares can amplify volatility. Investors may sell existing holdings to make room for incoming issuances or to raise cash for future quota requirements. The effect is often felt most strongly in the high-growth boards where valuations are most sensitive to liquidity shifts.
The artificial intelligence spending cycle
The AI spending pattern is not unique to China. Technology companies around the world are committing large sums to data centers, advanced chips, and model development. Many have yet to show meaningful revenue from those projects, and finance chiefs are being forced to explain why the outlays will eventually pay off. In China, however, the impact has arrived in reported earnings sooner and more visibly than in other large markets.
China is now the first major market to see a full capital expenditure cycle appear in reported profit and then be marked down by investors. The result is a cautionary example for other exchanges, because the same accounting arithmetic will show up elsewhere once heavy spending passes through income statements.
A different timetable in Europe
Europe sits at the opposite end of that cycle. Spending commitments have been made, but most have not yet reached the income statement, and much of the region’s exposure runs through the supply chain rather than through direct AI operators. The European Union has committed 20 billion euros to AI gigafactories. The programme has drawn 77 proposals across 16 member states and 60 sites, with construction of the first facility planned for 2027.
That timetable means European investors will have to wait years before seeing what AI spending truly does to corporate numbers. By the time the costs show up in profit reports, Shanghai’s markets will have already provided a live demonstration of how investors react when a capex cycle stops being a promise and becomes an expense line. For now, the lesson from China is that the highest growth can produce the sharpest reversal when enthusiasm outruns the financial reality.