Markets Close August on a High Note: AI Powerhouses Surge, Banks Hit Fresh Peaks, and Defense Stocks Extend Winning Streak

Deep News
3 hours ago

A-shares staged a remarkable recovery on the final trading day of August, opening lower before reversing course to close in positive territory. After a subdued morning session, all three major indices turned green in the afternoon, with heavyweight names like SMIC and Cambricon driving a significant rebound in the AI sector. Total market turnover reached 2.15 trillion yuan, with over 3,100 stocks advancing. Domestic computing power stocks delivered standout performances, as indices tracked by several tech-focused ETFs, including the STAR AI ETF (589520), Big Data ETF (516700), and Cloud Computing ETF (159099), all surged more than 3%.

According to CCTV Finance, the severe imbalance between domestic computing power supply and demand has pushed the industry chain from "product verification" into a full-scale "batch delivery" phase. Changxin Technology's first semi-annual report was explosive, posting revenue of 150.31 billion yuan—a year-on-year surge of 873.64%—and a net profit attributable to shareholders of 77.605 billion yuan, reversing a loss of 2.332 billion yuan in the same period last year. This financial validation of domestic storage capacity and profitability has lifted expectations across the entire supply chain. The STAR Chip ETF (589190), which holds nearly 50% of its assets in storage-related stocks, climbed from the day's lows to close with its benchmark index up 1.92%, after falling more than 2% earlier in the session.

In the banking sector, both BANK OF CHINA (601988) and CITIC BANK (601998) hit fresh all-time highs as the sector rallied broadly. Solid interim results from listed banks, combined with a series of new property policies including extended loan tenures, helped the Bank ETF (512800) benchmark rise 1.67%. Fund manager Feng Chencheng noted that bank revenues are expected to continue improving in the second half of the year, and with selling pressure on bank stocks significantly easing, the sector is well-positioned to outperform the broader market as risk appetite moderates and style rotation continues.

The defense and aerospace sector also attracted attention, with the Military ETF (512810) benchmark climbing 1.77% to secure a fifth consecutive day of gains. The State Council explicitly called for coordinated development of "space networks"—referring to space-air-ground integrated communication infrastructure led by low-orbit satellite internet. Overseas, SpaceX announced a $100 billion investment to build the world's largest launch site. Looking back at August, after July's pullback, all three major indices finished the month higher, with the electronics and communications sectors each gaining over 10% and non-ferrous metals leading the pack with an 11.35% rise. Huabao Fund suggests that while the hawkish tone from Jackson Hole may cause periodic market fluctuations, the non-ferrous metals sector—supported by tight supply-demand dynamics, low valuations, and clear demand driven by the intelligent industry trend—could offer buying opportunities on dips via the Non-Ferrous Metals ETF (159876).

Domestic Chips See Explosive Demand, Computing Power Needs 10x Supply

The AI sector saw a notable intraday reversal following official media reports of a severe supply shortage in domestic computing chips. Star Ring Technology led gains with an 11% surge, Opt Machine Vision rose over 10%, and Amlogic and Espressif each gained more than 8%. Cambricon and other major players also posted strong gains. The STAR AI ETF (589520), which allocates 42% of its holdings to GPU-related stocks, saw its benchmark index surge 3.72% to close at intraday highs.

The catalyst came from CCTV Finance reporting that domestic computing power demand exceeds supply by more than tenfold, with regular delivery times extending by a year and some high-end orders already scheduled three years out. As various smart computing complexes receive approval, the industry chain has fully transitioned to batch delivery. Biren Technology's revenue surged nearly 20-fold in its 2026 interim report, while Cambricon and Moore Threads also reported substantial revenue growth. Industry research indicates that demand for domestic AI chips in 2026 stands at approximately 4 million units, with actual delivery around 3 million—leaving a million-unit capacity gap. Industry consensus holds that domestic chips are in a strong upward cycle for high-end AI applications.

The domestic GPU industry has now completed a three-stage evolution: from "able to make" in 2023 (chip design and tape-out success), to "able to sell" (securing internet giant certifications and entering procurement lists with surging revenues), and now to "able to profit" (positive gross margins, positive non-GAAP net income, and sharply narrowing losses or returning to profitability). Guosheng Securities notes that leading chip design companies have already entered the profit phase. As domestic chip capabilities continue improving and narrowing the gap with overseas rivals, internet companies and smart computing centers are accelerating their adoption of domestic chips, creating a positive R&D cycle through expanded applications and accumulated experience. CICC points out that computing power has been elevated to national infrastructure status, with domestic large models iterating rapidly and "super nodes" shifting the competitive landscape from individual chip performance to system-level efficiency. The proven ability of domestic chips to handle real large model traffic means increased model usage will directly translate into higher utilization and procurement demand for domestic chips.

The STAR AI ETF (589520) and its feeder funds (A: 024560, C: 024561) focus on 30 large-cap STAR Market companies providing foundational resources, technology, and application support for AI, with semiconductor exposure at 70.4%. The fund offers strong offensive characteristics with GPU concept stocks at 41.98% and AI application stocks at 23.19% weightings. With 20% daily price limits, the ETF provides low-threshold access to the STAR Market's breakthrough potential and serves as a margin trading target for efficient domestic computing power allocation.

AI Applications Reach "Moon Landing" Moment: New Drama Release Sparks Limit-Up Rally

AI applications took center stage as the first AI-produced TV drama to air on satellite television triggered a wave of limit-up moves in related stocks. The ChiNext AI ETF (159363) benchmark surged over 2% in afternoon trading, maintaining high market attention. Mango Excellent Media and China Online both jumped 20% to their daily limits, Huace Film & TV gained over 15%, SigmaStar Technology rose more than 12%, and WPS, Allwinner Technology, Kunlun Tech, Sangfor Technologies, and Yeahka all advanced over 7%. Optical module and CPO stocks were relatively subdued, with Zhongji Innolight and TFC Communication closing slightly lower.

The catalyst was the announcement that "Journey to the West: A New Tale," produced by Mango TV and hailed as China's first AI-produced drama to air on satellite, premiered on August 31. The first five episodes of the "Huaguoshan" arc aired at 6 PM on Mango TV and 8 PM on Hunan Satellite TV. Fully produced by AIGC with no human actors, the series marks the first "produce-while-airing" production following the "Radio and Television 21 Measures." With 30 episodes scheduled for prime time, the drama signals AIGC's transition from demo stage to industrialized production. China Merchants Securities points out that AI short dramas have a clear business model suitable for rapid domestic growth and have already expanded successfully overseas. As the industry consolidates, overseas markets open up, and content quality continues to improve, AI application leaders are expected to maintain their momentum. The brokerage remains bullish on AI short dramas and the broader AI film/TV production chain.

Regarding optical modules, Guosheng Securities notes that the two leading manufacturers, Zhongji Innolight and Eoptolink, have a combined potential deliverable module inventory exceeding 28 million units, providing solid material support for continued shipments through next year and validating delivery certainty during this period of high industry prosperity. The firm continues to recommend the communication sector's interim report opportunities and related computing power chain companies. Investors seeking exposure to the high-"light" theme and AI applications may consider the ChiNext AI ETF (159363) and its OTC feeder funds (A: 023407, C: 023408), which focus on optical module and CPO leaders while balancing AI application exposure. The benchmark index holds nearly 40% combined weight in Zhongji Innolight, Eoptolink, and TFC Communication.

SpaceX's $100 Billion Investment Boosts Defense Stocks to Five-Day Winning Streak

The defense sector powered higher on August 31, with the CSI Military Industry Index climbing 1.77% to secure a fifth consecutive day of gains. Military electronics and aerospace concepts led the advance: Raytron Technology and Bright Laser Technologies each gained over 14%, Aerospace Nanhu rose 7.6%, and Sichuan Jiuzhou Electric and Feilihua advanced more than 6%. The Military ETF (512810), which passively tracks the index, opened lower but finished near session highs with approximately 3% intraday volatility.

August saw the defense sector stabilize after a period of consolidation. Two key drivers are emerging. First, event catalysts: the State Council recently called for the first time for coordinated development of "space networks" (low-orbit satellite internet and space-air-ground integrated communications infrastructure). Internationally, on August 25, SpaceX announced a $100 billion investment to construct the world's largest launch site. Second, interim earnings: the 80 defense leaders covered by the Military ETF (512810) posted combined first-half revenue of 377.494 billion yuan, up 18.89% year-on-year, with net profits attributable to shareholders of 24.772 billion yuan, up 32.92%. The overall improving performance of core defense listed companies validates the sector's upward trajectory. Looking ahead, Orient Securities expects the "15th Five-Year Plan" to be finalized before year-end, with catalysts including the Zhuhai Airshow in Q4. At current levels, defense blue chips may be in a left-side accumulation phase, with a potential right-side opportunity emerging in Q4 as the trend turns upward.

The Military ETF (512810), formerly the National Defense ETF, passively tracks the CSI Military Industry Index, offering comprehensive coverage of commercial aerospace, low-altitude economy, large aircraft, MLCC, military AI, and gas turbine themes. As a margin trading and Stock Connect-eligible instrument, it serves as an efficient core asset vehicle for defense investment.

Source: Shanghai and Shenzhen exchanges, as of August 31, 2026. Note: Fund rates are detailed in respective fund legal documents. The "storage content" refers to the Storage Chip Index (980138.CNI) weighting of 49.95% as of August 26. GPU and AI application concept weightings compare against the GPU Index (8841701.WI) and AI Application Index (980112.CNI). As of August 21, 2026, the top three constituents of the ChiNext AI Index were Eoptolink (13.77%), Zhongji Innolight (13.17%), and TFC Communication (10.56%), according to Guozheng Index. Institutional views reference: 1) Guosheng Securities' late-August 2026 computing/semiconductor industry research; 2) CICC's August 27-31 computing power industry special report; 3) China Merchants Securities on AI short dramas as leaders in AI applications; 4) Guosheng Securities on optical module delivery reserves from interim reports; 5) Orient Securities' defense industry weekly report from August 24. Risk disclosure: The STAR AI ETF passively tracks the STAR AI Index (base date December 30, 2022; published July 25, 2024). The ChiNext AI ETF tracks the ChiNext AI Index (base date December 28, 2018; published July 11, 2024), with annual returns of 17.57%, -34.52%, 47.83%, 38.44%, and 106.35% from 2021-2025 and corresponding annualized volatility of 23.73%, 27.34%, 38.02%, 45.42%, and 41.1%. The Military ETF tracks the CSI Military Industry Index (base date December 31, 2004; published December 26, 2013), with annual returns/volatility of 14.28%/33.05%, -25.74%/23.44%, -11.02%/18.34%, 8.20%/34.39%, and 31.55%/21.43% from 2021-2025. Index constituents adjust per index methodology; backtested performance does not indicate future results. Mentioned stocks are for index constituent demonstration only and do not constitute recommendations. All information herein is for reference only; investors bear responsibility for their own investment decisions. This content does not constitute investment advice. Investors should review fund contracts, prospectuses, and product fact sheets to understand risk-return profiles. Past performance does not guarantee future results. The STAR AI ETF (589520) and ChiNext AI ETF (159363) are rated R4 (medium-high risk), suitable for aggressive investors (C4 and above); the Military ETF (512810) is rated R3 (medium risk), suitable for balanced investors (C3 and above). Suitability opinions may vary by distributor. Fund risk ratings from distributors may differ from those of fund managers. Investors should consider their own objectives, time horizon, experience, and risk tolerance. CSRC registration does not constitute an endorsement of investment value or returns. Fund investment requires caution.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Most Discussed

  1. 1
     
     
     
     
  2. 2
     
     
     
     
  3. 3
     
     
     
     
  4. 4
     
     
     
     
  5. 5
     
     
     
     
  6. 6
     
     
     
     
  7. 7
     
     
     
     
  8. 8
     
     
     
     
  9. 9
     
     
     
     
  10. 10