China’s listed companies produced their strongest profit growth in years during the second quarter, but the results were not enough to sustain a market rally in AI-related stocks. Bloomberg reports that a China International Capital Corp. assessment found profits at onshore-listed A-share companies rose 25.7% year over year in the second quarter of 2026, the fastest pace in almost five years.
The strongest growth came from sectors associated with computing and AI infrastructure. The CICC research cited by Bloomberg said the constituents of the CSI Artificial Intelligence Index posted 95% profit growth in the quarter. Profits in the technology, media, and telecommunications semiconductor segment rose 202%, and earnings at STAR Market chip companies rose 234% from a year earlier.
Those figures are CICC research estimates and sector aggregates, not a report that every AI or chip company is growing quickly. They also cover second-quarter financial results, while the market reaction reflects investors’ view of the future. Bloomberg said the CSI 300 Index had fallen about 9% in the third quarter, and the technology-heavy STAR 50 had dropped 29% despite the apparent earnings strength.
AI Hardware Profits Are Rising Faster Than Broad Corporate Earnings
The 25.7% increase in A-share profits sets the broad context, but the CICC figures show a much sharper divide among sectors. The CSI AI Index’s reported 95% gain was driven by computing-power demand, according to the Bloomberg account. In the same research, optical-module profits rose 140% year over year and printed-circuit-board earnings rose 98%.
These are supply-chain categories that can benefit when cloud operators and hardware makers add servers, networking equipment, and related systems. Optical modules carry data across high-speed networks. Printed circuit boards are essential for servers, communication devices, and other electronic equipment. Their growth does not prove that every AI deployment is profitable, but it shows that demand for infrastructure can support particular manufacturing segments.
CICC also reported that 40% of industries still experienced year-over-year profit declines. That figure is a reminder that strong performance in semiconductors and AI hardware did not spread evenly through the economy. Investors can see a visible AI spending cycle while remaining concerned about consumer demand, property conditions, export uncertainty, or the ability of companies to turn computing investment into durable returns.
The earnings data build on a market story EastFrontier covered earlier in the week. On September 2, foreign funds turned toward China’s AI hardware stocks, showing how demand for chips, servers, and related equipment had attracted attention from global investors. Bloomberg’s new report explains why that attention can coexist with a broad pullback: strong company earnings are not always enough when valuations and macroeconomic expectations have already moved higher.
The STAR 50 Decline Reflects a Debate About Returns
A 29% drop in the STAR 50 is particularly relevant because the index is home to many technology and semiconductor companies that investors associate with China’s innovation strategy. The fall does not erase the CICC earnings figures. It suggests that investors are asking whether current growth can continue, whether margins will remain strong, and whether the market has already priced in too much AI optimism.
The same tension has appeared in other markets, where investors distinguish between companies selling infrastructure and companies trying to monetize AI applications. Chip makers, optical-module suppliers, and board manufacturers can benefit quickly when capital spending increases. The longer-term question is whether cloud operators and enterprise customers earn enough from AI services to keep that spending pace intact.
CICC’s sector breakdown shows another contrast. It reported that upstream energy and raw-material industries achieved 74.6% profit growth, compared with 12.4% for downstream sectors. The difference indicates that the earnings revival was not limited to AI. Commodity prices, energy conditions, and industrial demand also shaped the quarter’s results.
The research further said that renminbi appreciation generated 107 billion yuan in foreign-exchange gains for nonfinancial companies during the first half of 2026, amounting to 5.5% of total net profit. That is another reason to avoid treating all of the A-share earnings growth as an AI effect. Exchange-rate gains can boost reported profits even when operating conditions are uneven.
EastFrontier reported in August that China’s AI boom had entered a tougher phase as startups raced toward IPOs. The Bloomberg and CICC figures extend that question from venture-backed companies to listed corporations. The market is rewarding visible hardware demand in some cases, but it is also demanding evidence that the broader AI investment cycle will produce lasting returns.
Earnings Strength Does Not Eliminate Valuation Risk
Investors often use earnings growth as a reason to support higher valuations. The recent pullback shows why the relationship is not automatic. A company can report strong quarterly profits and still see its share price fall if investors think future growth will slow, competition will reduce margins, or the market had expected even better results.
For China’s AI sector, that issue is especially acute because the investment cycle has multiple layers. One group of companies sells chips, memory, optical modules, boards, data-center equipment, and power systems. Another builds models, cloud services, agents, or enterprise applications. A third group consists of customers that must decide whether AI saves enough money or creates enough revenue to justify continued spending.
CICC’s 202% semiconductor-profit growth and 234% STAR chip-company increase suggest that the first layer is currently strong. The decline in the broader CSI 300 and STAR 50 suggests the market is less certain about the second and third layers. The data do not demonstrate an AI bubble, nor do they prove a durable earnings boom. They show that investors are differentiating more carefully between current supplier results and future application economics.
The same caution applies to the CICC numbers themselves. They are an analysis of listed-company financial disclosures, and the source does not provide a complete company-level table in the Bloomberg article. Readers should use the percentages as an indicator of sector trends rather than as a substitute for examining individual balance sheets, orders, margins, and cash flow.
China’s AI hardware supply chain has a real earnings story in the second quarter. The market reaction shows that investors want more than a strong quarter. They want evidence that computing demand, enterprise adoption, and profitability can remain robust after the initial surge. The next earnings season will show whether the gap between company results and share-price performance narrows or becomes a more persistent feature of China’s AI market.