The domestic AI industrial chain experienced an afternoon surge today (August 31st), with StarRing Technology leading the rally with gains exceeding 10%, while Opt Machine Vision climbed 8% and Espressif Systems advanced over 7%. Other notable advancers included Amlogic and Fudan Microelectronics. The benchmark index tracked by the HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC (589520), an ETF with a 42% GPU concentration, rose more than 3% intraday. The chart above illustrates the top 10 performing constituent stocks of this ETF on August 31st.
On the news front, on August 26th, Zhipu announced the launch and open-sourcing of the GLM-5.3-Flash model, with its public services entirely powered by domestic computing infrastructure. This included seamless adaptation by Cambricon Technologies Corporation Limited on the same day, validating the feasibility of domestic chip solutions. The end-to-end service performance tripled compared to the same hardware baseline, achieving hardware efficiency and per-token costs on par with mainstream NVIDIA GPUs. On the earnings front, during the first half of 2026, leading domestic computing chip makers Cambricon Technologies Corporation Limited and Hygon Information Technology both reported rapid growth in revenue and profits, reflecting the increasing adoption of domestic chips amid sustained high demand for AI computing power.
From a technological perspective, the primary challenge for domestic computing power is shifting from "chip shortages" to "maximizing chip efficiency." As large model computational complexity continues to rise, the efficiency bottlenecks of traditional computing clusters are becoming more apparent. In the general-purpose computing era, CPU-centric architectures with limited node collaboration relied mainly on standalone machines or small clusters to meet fundamental computing needs. With the expansion of deep learning models and the adoption of the Transformer architecture, GPUs have become the core of computing power, with systems evolving into multi-GPU Scale-out cluster architectures.
Where to begin — According to Guosheng Securities, leading chip design companies have already entered a profitable phase. The continuous improvement in domestic chip capabilities is narrowing the gap with overseas counterparts, accelerating the adoption of domestic chips in internet and computing center deployments. Domestic computing demand is increasingly shifting to local manufacturers, and expanded application scenarios coupled with accumulated experience are driving iterative chip improvements, gradually forming a positive R&D loop. Guolian Minsheng Securities added that computing network construction is entering an accelerated implementation phase, driven by trillion-yuan level direct investments planned for the 15th Five-Year Plan period. Combined with robust earnings growth at domestic computing companies and sustained increases in capital expenditure from major tech firms, the domestic computing power sector is expected to maintain strong momentum. Furthermore, the urgency and strategic importance of developing indigenous large models and computing power are becoming increasingly pronounced, presenting significant strategic opportunities for the synergy of domestic models and chips.
Only a handful of key players matter, with the HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC (589520) and its feeder funds (Class A: 024560, Class C: 024561) focusing on the domestic AI industrial chain. These funds hold 30 large-cap STAR Market-listed companies that provide foundational resources, technology, and application support for AI. Semiconductor stocks account for 70.4% of the portfolio, offering strong offensive characteristics, with GPU concept stocks and AI application concept stocks making up 41.98% and 23.19%, respectively. With its 20% daily price limit and low entry threshold, this ETF serves as an efficient vehicle for accessing growth opportunities in the science and technology innovation sector. Additionally, as a margin trading and securities lending target, it represents a convenient tool for one-stop investment in domestic computing power. It is important to note that GPU and AI application concept weightings are compared against the GPU Index and AI Application Index, as sourced from exchanges up to August 31, 2026.
Regarding fees, the ETF does not charge sales service fees, with authorized brokers permitted to collect commissions up to 0.5%, which includes fees charged by stock exchanges and registration institutions. On-exchange trading fees are subject to actual charges by securities firms. Risk disclosure: HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC passively tracks the SSE STAR Market Artificial Intelligence Index, with a base date of December 30, 2022, and a release date of July 25, 2024. The index's constituent stocks are adjusted according to its compilation rules, and backtested historical performance does not guarantee future results. The stocks and index constituents mentioned here are for illustrative purposes only and do not constitute investment advice or reflect portfolio holdings or trading activities of any fund under management. The fund manager rates this ETF as R4 (moderate-to-high risk), suitable for aggressive (C4) and above investors, with suitability recommendations confirmed by distribution institutions. Any information presented herein (including but not limited to stocks, commentary, forecasts, charts, indicators, theories, and any other forms of expression) is for reference only, and investors must bear full responsibility for their own investment decisions. Moreover, any views, analyses, or forecasts in this article do not constitute investment advice to readers and shall not incur liability for direct or indirect losses arising from the use of this content. Fund investments carry risks; past performance does not guarantee future results, and the performance of other funds managed by the same manager should not be relied upon as an indicator of this fund's future performance. Investment in funds requires caution.
Comments