On October 10, Wu Qing, Secretary of the Party Committee and Chairman of the China Securities Regulatory Commission, held a symposium in Beijing with responsible persons of certain listed companies and experts from securities fund institutions to deeply exchange views and fully solicit opinions and suggestions on the current capital market and financial situation.
At the symposium, participants unanimously agreed that China's macroeconomy is stable and improving, industrial structure transformation continues to deepen, the capital market operates generally steadily, and the functions of the multi-tiered market are effectively brought into play. Although facing some risks and challenges, the trend of high-quality development will not change, and there is full confidence in economic operation and the capital market.
Wu Qing stated that listed companies and industry institutions are important entities in the market, and he hopes everyone will actively offer suggestions and jointly build the capital market well, better repay the vast number of investors, and actively contribute to the advancement of Chinese-style modernization.
On October 9, the "Opinions on Developing New Quality Productive Forces" (hereinafter referred to as the "Opinions") was issued.
The "Opinions" places "vigorously promoting scientific and technological innovation" as the first of five major tasks, and further deploys "strengthening the dominant position of enterprises in scientific and technological innovation."
Among them, it points out the comprehensive implementation of the "AI+" action to advance the transformation of traditional industries by artificial intelligence.
R&D projects of enterprises that meet national strategic needs will be elevated to national projects according to procedures, enterprises will be encouraged to lead national scientific and technological tasks, and capable private enterprises will be supported to lead major national technological breakthrough tasks.
CICC: How Will A-Share Market Style Play Out? After a Phase of Highly Concentrated Trading, Small-Cap Style May Have Relative Advantages
From June to August 2026, the trading heat of the STAR Market rose periodically, while the trading heat of the Dividend Index remained relatively sluggish.
The STAR 50 Index, centered on hard technology such as semiconductors and communications, saw both prices and trading volume rise simultaneously from June to August 2026; this round of incremental funds did not flow evenly across the entire market, but concentrated into a small number of growth directions such as semiconductors and communications.
Specifically, from June to August, the stocks that were concentrated in trading were mainly in the AI hardware sector, contributing greatly to the index weights of the STAR and ChiNext boards.
The electronics industry contributed about 36% of the high trading records, noticeably higher than other industries; 28 constituent stocks of the STAR 50 belonged to the concentrated trading stocks, with a combined weight of about 85%; ChiNext Growth had 41 such stocks, with a weight of about 96%; the weight proportions of the ChiNext Index and CSI 300 were also about 88% and 77%, respectively.
After the phase of highly concentrated trading, the concentrated trading stock portfolio continued to underperform the CSI 1000 Index; the sample statistics covered the excess return performance of the concentrated trading stock portfolio relative to the CSI 1000 and CSI Dividend over a period after concentrated trading receded.
It can be found that the concentrated trading stock portfolio had a relatively low win rate relative to the CSI 1000 Index during this period, and the longer the holding period, the lower the win rate; although the win rate relative to the CSI Dividend Index was also relatively low, as the holding period lengthened, the win rate began to rise.
From the perspective of index return performance, after the phase of highly concentrated trading, the win rate of CSI 1000 relative to CSI 300 exceeded 90%.
Using trading days with concentration above 46% as samples, the excess win rate of CSI 1000 relative to CSI 300 rose from about 64% in the 10-day window to about 96% in the 100-day window, with windows above 60 days all approaching or exceeding 90%.
Shanxi Securities: The "15th Five-Year Plan" for the New Battery Industry Is the First National-Level Plan for the Battery Industry, with Industry Scale Growing Steadily Through 2030
On September 28, seven departments jointly issued the "15th Five-Year Plan for the Development of the New Battery Industry" (hereinafter referred to as the "Plan"), which is China's first national-level special five-year plan in the battery field.
The "Plan" proposes that by 2030, the industry scale will grow steadily, all-solid-state batteries will initially achieve large-scale application, the cycle life of long-life lithium batteries will reach 15,000 cycles, and the product defect rate of leading enterprises will reach the PPB level. It also deploys 19 key tasks and 5 special columns from five major aspects: full-chain innovation, industrial transformation and upgrading, industrial system construction, science and technology services, and trade and investment cooperation.
The historical upgrade of the industry's status is unfolding, and the application scenario map for new batteries is gradually spreading out.
In the past, new batteries were long regarded as "components" of the automotive and electronics industries. This plan positions them for the first time as an underlying basic industry for building a new energy system and cultivating new quality productive forces, and as the development foundation for emerging tracks such as energy storage, low-altitude economy, and humanoid robots.
The upgrade in positioning means that in the future, policy resources for all downstream tracks (special bonds, scenario opening, procurement support) will give batteries "ride-along" eligibility. This policy spillover value is far greater than the plan text itself.
The list of seven departments includes the National Railway Administration, which is rare, and is itself an application scenario map.
Supply governance and demand expansion are being pursued simultaneously.
The "Plan" clearly states "implementing capacity warning and regulation, standardizing price competition, and optimizing regional layout," which is a direct response to the current contradiction in the supply-demand structure of lithium batteries, equivalent to a battery-sector "anti-involution," with a five-year institutional lever for the clearing path.
On the demand side, energy storage scenarios are extended to intelligent computing centers, green power direct connection, virtual power plants, and computing-power coordination, and "AI power shortage" has officially become the second growth curve for battery demand.
In terms of investment allocation, under policy promotion, the iteration of new battery technologies is expected to accelerate, helping clear backward production capacity in the industry, quickly break away from disorderly involutionary competition, and usher in new growth opportunities for the industry. Attention may be paid to Lingpai Technology, Putailai, Tinci Material Technology, Sige New Energy, and others.
Jinyuan Securities: Breakthroughs in Reusable Technology for Commercial Aerospace, Opening a Cost Revolution in Space Launch Capacity
Commercial aerospace is a market-oriented industrial chain of "satellite-rocket-site-testing-application."
Frequency and orbit rules turn long-term plans into time-limited deployment tasks.
The ITU sets activation and phased deployment requirements for non-geostationary orbit satellite systems in applicable frequency bands: after activation is completed, they must reach 10%, 50%, and 100% of the declared scale in the 2nd, 5th, and 7th years thereafter, respectively.
Therefore, already-declared constellations must not only "launch the first satellite," but also continuously complete deployment milestones.
The defense system is shifting from a small number of high-value satellites to distributed, iterable constellations.
The U.S. Space Development Agency's PWSA adopts a rolling deployment model with one generation every two years; Tranche 1 plans 154 combat satellites and 4 demonstration satellites, completing constellation formation through continuous launches.
Domestic commercial launches have entered a high-frequency stage.
In 2025, China completed 50 commercial aerospace launches, including 25 commercial launch vehicle launches; 311 commercial satellites entered orbit for the year.
Scenarios such as satellite communications, remote sensing mapping, navigation augmentation, military reconnaissance, and emergency rescue together form the pool of launch capacity demand.
Demand judgment should shift from the "number of declared satellites" to executable indicators.
The reusable architecture turns the first-stage engine, tanks, rocket body, and some electronic systems from one-time consumables into revolving assets, but at the same time increases return propellant, recovery facilities, inspection, and refurbishment costs. Actual economics depend on the number of reuses and turnover efficiency.
The product of commercial launch is not just the rocket itself.
Complete delivery also includes mission analysis, satellite-rocket matching, fairing environment, test-launch control, flight safety, orbital injection accuracy, and launch window assurance; reliability, punctuality, and mission adaptability together determine service value.
In 2025, the core industry scale of China's commercial aerospace reached 1.01 trillion yuan, a year-on-year increase of nearly 7%.
Innovation remained active in links such as satellites, rockets, launch sites, TT&C, and applications, opening broader space for industry development.
According to forecasts by CCID Think Tank, during the "15th Five-Year Plan" period, the compound growth rate of China's commercial aerospace core industry scale will reach about 11%, maintaining a rapid development trend.
Related listed companies include Aerospace Power, BLT (rights protection), AVIC Aerospace Technology, Haohua Technology, and others.
Sinolink Securities: The Next Stop for AI, Physical AI
Physical AI represented by world models can be regarded as an intermediate station from LLMs to superintelligent robots.
LLMs excel at language understanding, knowledge retrieval, and task decomposition, and have already achieved commercialization loops of considerable scale in white-collar work fields such as coding, general office work, finance, and law, though they are still at an early penetration stage.
We believe that the next stop for AI is entering the physical world. Physical AI also needs to understand three-dimensional space, object states, contact relationships, and dynamic constraints, and predict the possible results of different actions.
AMD's acquisition of World Labs pushes world models into an industrial investment phase, further deepening the synergy between models and computing platforms.
On September 28, AMD announced a planned all-stock acquisition of World Labs for about US$8.2 billion. The transaction is expected to be completed by the end of 2026 and still needs to satisfy closing conditions such as regulatory approval; Atlas, released by World Labs, can natively process text, images, video, camera poses, and 3D information.
Current technology has roughly formed four paradigms: 3D space generation, pixel and video prediction, latent variable prediction, and physics- and action-conditioned models. Evaluation standards have also extended from image quality to spatial consistency, physical plausibility, action controllability, and long-term prediction capability.
Scenarios, data, and models for embodied intelligence are evolving simultaneously, and high-value, low-repetition tasks are expected to form commercial loops first.
Standardized, highly repetitive tasks are already covered by general or specialized equipment. Non-standard tasks such as disaster rescue, complex equipment maintenance, flexible manufacturing, and home care better reflect the cross-scenario value of humanoid robots; the industry shows early characteristics of a "GPT-2 moment," but the success rate of some publicly disclosed multi-fingered dexterous tasks remains at 32% to 44%. Generalization in unfamiliar environments, long-task stability, and the frequency of human intervention remain core constraints on large-scale deployment.
In terms of investment allocation, attention may be paid to the North American chain, including Kersen Technology, Siling Intelligent Drive, Forsai Technology, and others; other robot chains include Lingyi iTech, Pan Asia Micropenetration, Orbbec, and others.