On February 12, Zhipu AI unveiled its new flagship model, GLM-5. Previously circulated under the anonymous codename "PonyAlpha" on overseas platforms, the model ignited discussions within the tech community. It has been hailed internationally as the "most powerful open-source code model," with reasoning depth and engineering capabilities approaching top-tier closed-source standards. This development potentially signifies China's official entry into the global first tier in areas such as sparse architecture, coding ability, and intelligent agents.
When an open-source model begins to approach the level of leading closed-source models in long-range task planning and autonomous error correction, the foundational logic of the specialization chain begins to weaken. Industry transformation often occurs not at the moment of "surpassing," but at the moment of becoming "sufficiently capable." Once the capability ceiling of open-source models reaches the threshold of closed-source ones, the pricing power of the closed-source camp starts to face pressure.
It is precisely this shift in the underlying business logic that has caught Wall Street's attention, turning its gaze eastward. The unprecedented high valuations and core strategic positioning that global top-tier investment banks are assigning to Chinese large language model companies essentially signal, with real capital, that the market is reassessing the value anchors of China's foundational AI enterprises.
Consequently, technological breakthroughs, capital revaluation, and top-level policy design have converged historically during the 2026 Spring Festival. On the policy front, Zhipu AI founder Tang Jie recently reported to senior leadership as a representative of major model enterprises. In the same week, the State Council held a special study session, explicitly emphasizing the need to "deepen and expand 'AI+' to empower all industries comprehensively."
Industry insiders point out that if the 2025 Spring Festival was DeepSeek's "lone warrior moment," the 2026 Spring Festival presents a different picture for Chinese AI—no longer characterized by breakthroughs from a single company, but rather a collective advance on three fronts: visual, engineering, and foundational models. The Chinese AI sector has essentially completed a strategic deployment akin to the "Two Bombs, One Satellite" project for the digital age:
1. Seedance 2.0 is the "visual bomb" that shattered perceptual limits. It led overseas practitioners to exclaim that it is "two generations ahead," making "the end of AIGC's childhood" an industry consensus. 2. GLM-5 is the "engineering bomb" that penetrated the deep waters of productivity. It caused overseas developers to mistake this Chinese open-source model for a top-tier closed-source product from Silicon Valley, creating, for the first time in the history of domestic models, a seller's market where demand outstrips supply. 3. DeepSeek is the high-altitude "ecosystem satellite." Its core technology is providing underlying navigation for the entire Chinese AI ecosystem through technological spillover, fostering powerful interoperability between enterprises.
Looking ahead, major model releases such as ByteDance's Doubao 2.0, Alibaba's Qwen 3.5, and DeepSeek V4 are imminent, potentially heralding the next wave of AI advancement.
In market performance today (February 13), the HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC (589520), which focuses on the domestic AI industry chain, saw its intraday price rise against the market trend by over 1.7% before being pulled down by the broader market. It is currently up 0.29%, with its daily K-line potentially forming an ascending staircase pattern.
Among its constituent stocks, Transwarp Technology led gains, rising over 6%. AsiaInfo Technology gained more than 4%, while QiAnXin Technology, Anlogic, and VeriSilicon Microelectronics each rose over 3%. Stocks such as Scantech, Opt, and Allwinner Technology also followed with increases.
[A Beacon of Domestic Substitution, Strengthening Sci-Tech Self-Reliance] The HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC (589520) and its feeder funds (Feeder A: 024560, Feeder C: 024561) primarily invest in the domestic AI industry chain. Their constituent stocks include leading domestic GPU companies (e.g., Cambricon), leading domestic ASIC companies (e.g., VeriSilicon Microelectronics), and leading AI application companies (e.g., Kingsoft Office). The semiconductor sector carries a weight of nearly half, lending the fund strong offensive potential. The software sector, with a weight exceeding 30%, stands to benefit from potential catch-up rallies in AI applications. Furthermore, this ETF is a margin trading security, making it an efficient tool for gaining exposure to domestic computing power.
ETF Fee Information: The HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC does not charge a sales service fee. Subscription and redemption agents may charge a commission of up to 0.5%, which includes fees levied by stock exchanges and registration institutions. On-market trading fees are subject to the rates charged by the securities firm.
Feeder Fund Fee Information: For the HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC Feeder Fund (Class A), the subscription fee rate is 1,000 RMB per transaction for subscription amounts of 2 million RMB or more; 0.6% for amounts between 1 million RMB (inclusive) and 2 million RMB; and 1% for amounts below 1 million RMB. The redemption fee rate is 1.5% for holding periods under 7 days, and 0% for holding periods of 7 days or more. No sales service fee is charged. The HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC Feeder Fund (Class C) charges no subscription fee. The redemption fee rate is 1.5% for holding periods under 7 days, and 0% for holding periods of 7 days or more. A sales service fee of 0.3% is charged.
Risk Warning: The HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC passively tracks the SSE Science and Technology Innovation Board Artificial Intelligence Index. The base date for this index is December 30, 2022, and it was published on July 25, 2024. The index's annual gains/losses for 2023 and 2024 were 12.68% and 32.36% respectively. The composition of the index's constituent stocks is adjusted according to its compilation rules, and its past performance does not indicate future returns. Individual stocks and index constituents mentioned herein are for illustrative purposes only; descriptions of individual stocks are not investment advice of any form and do not represent the holdings or trading动向 of any fund managed by the asset manager. The fund manager assesses the risk rating of the HUABAO SHANGHAI SCIENCE AND TECHNOLOGY INNOVATION BOARD ARTIFICIAL INTELLIGENCE TRADING OPEN ENDED INDEX SEC as R4 - Medium-High Risk, suitable for aggressive (C4) and higher risk-profile investors. The appropriateness matching opinion is subject to the final assessment by the selling institution. Any information appearing in this article is provided for reference only, and investors are solely responsible for any independent investment decisions. Furthermore, any views, analysis, or predictions herein do not constitute investment advice to readers and shall not be liable for any direct or indirect losses resulting from the use of this content. Fund investment carries risks; past performance of a fund does not guarantee future results, and the performance of other funds managed by the fund manager does not constitute a guarantee of the fund's performance. Invest cautiously in funds.
A MACD golden cross signal has formed, and these stocks are performing well.
Comments