In September 2026, the AI concept stock market entered a phase of divergence and adjustment after a rapid rise in the first half of the year. As of September 28th, the AI concept stock index (hereinafter referred to as "AI Index") closed at 12,580 points, a 2.3% decrease from the end of August, yet still up 15.6% from the beginning of the year, indicating that the market's long-term confidence in the AI sector remains unchanged. However, structural divergence within the sector has intensified, with the trends of computing power and applications showing stark differences: the computing power sector (such as chips and servers) declined by an average of 4.1%, while the application end (such as AI+Healthcare, AI+Finance, and B-side software) rose by 1.8%, becoming a new market hotspot. I. Market Overview: AI Concept Stocks Show Divergence, with Applications as a Safe Haven As September began, the AI concept stock market presented a "two extremes" pattern, influenced by multiple factors. On the one hand, the computing power sector remained under pressure due to factors such as tight chip supply and high costs, with leading companies like SMIC and Cambricon experiencing stock price pullbacks. On the other hand, the application end became a focus of capital inflow due to accelerated commercialization. For example, individual stocks in the AI+Healthcare sector, such as Huikang, and in the AI+Finance sector, such as Tonghuashun, all saw gains exceeding 5% in September, far surpassing the market average. This divergence is an inevitable result of the AI industry transitioning from "technology-driven" to "business-driven." In 2026, the AI industry has entered a critical stage of commercialization, with the development of the application end being jointly promoted by policy support (such as the Ministry of Industry and Information Technology's "AI Industry Development Plan for 2028") and market demand (such as enterprise digital transformation needs). Computing power, as the "infrastructure" of AI, although having a certain long-term demand, faces short-term issues such as overcapacity and cost pressure, leading to temporary capital withdrawal. II. Main Trend Analysis: Divergence Between Computing Power and Applications, with Commercialization as the Key 1. Computing Power Sector: Short-Term Pressure, Long-Term Demand Remains The computing power sector is a traditional main trend for AI concept stocks, covering segments such as chips, servers, and data centers. In September, this sector performed weakly, mainly due to the following factors: * **Easing Chip Supply Tensions, Highlighting Cost Pressure**: In 2026, as manufacturers like TSMC and SMIC expand capacity, the tight chip supply situation has eased, but high-end chips (such as AI training chips) remain in short supply. At the same time, the cost of chip manufacturing has increased, leading to high prices for computing power equipment and decreased corporate purchasing willingness. * **Adjustment in Capital Expenditure Cycle**: Global AI capital expenditure peaked in 2025 and entered a "mid-term rest" phase in 2026. Institutions estimate that global AI capital expenditure in 2026 will be 15% year-on-year, a significant decrease from the 30% in 2025, leading to a short-term slowdown in computing power demand. * **Intensifying Competition**: Domestic computing power manufacturers such as Huawei and Inspur are increasing investment, leading to intensified market competition and compressed profit margins. Despite short-term pressure, the long-term demand for the computing power sector remains strong. According to data from the Ministry of Industry and Information Technology, the scale of the core AI industry will reach 800 billion RMB by 2028, with the computing power industry accounting for approximately 40%, or 320 billion RMB. With the popularization of AI applications, computing power demand will continue to grow, retaining investment value in the long run. 2. Application End: Accelerated Commercialization, Becoming a New Favorite for Capital The application sector is the new main trend for AI concept stocks, covering fields such as AI+Healthcare, AI+Finance, B2B Software, and AI+Manufacturing. In September, this sector experienced an upward trend against the general market, mainly due to the acceleration of commercialization:
  • AI+Healthcare: Policy Support and Demand Growth: In 2026, the National Health Commission released the "AI Medical Application Guidelines," encouraging the application of AI technology in healthcare. Concurrently, the increasing demand for hospital digital transformation has accelerated the implementation of applications like AI-assisted diagnosis and intelligent consultation, driving revenue growth for related enterprises.
  • AI+Finance: Compliance and Efficiency Improvement: The financial industry's demand for AI is mainly concentrated in areas such as compliance risk control and intelligent advisory. In 2026, the central bank issued the "Financial AI Application Regulatory Measures," standardizing AI use in finance while improving the operational efficiency of financial institutions, thereby promoting performance improvement for AI+Finance companies.
  • B2B Software: Enterprise Digital Transformation Needs: With the deepening of enterprise digital transformation, the demand for AI in B2B software (such as ERP and CRM) has increased. For example, enterprises like Yonyou Network and Kingdee International launched AI+ERP products to help companies improve management efficiency, which was well-received by the market.
  • AI+Manufacturing: Smart Manufacturing Upgrade: Manufacturing is a key area for AI application. In 2026, the Ministry of Industry and Information Technology promoted the "Smart Manufacturing 2025" plan, encouraging enterprises to adopt AI technology to enhance production efficiency. For instance, the implementation of applications like industrial robots and intelligent quality inspection has driven revenue growth for related companies.
The commercialization achievements in the application sector have given this segment higher certainty and growth potential, making it a favorite among institutional funds. For example, public funds significantly increased their holdings of application-side stocks like AI+Healthcare and AI+Finance in September, while reducing holdings in computing power stocks.

III. Capital Flows: Institutional Reallocation from Computing Power to Application Sector

In September, the capital flow of AI concept stocks changed significantly, with institutional funds shifting from the computing power sector to the application sector. According to Wind data, in September, the proportion of the computing power sector in public funds' holdings of AI concept stocks decreased from 35% in August to 30%, while the proportion of the application sector increased from 25% to 32%. This reallocation logic stems from the institutions' focus on "commercialization achievements." Application-side companies, supported by clear performance backing (such as revenue growth in AI+Healthcare enterprises), were favored, while computing power companies faced short-term performance pressure and were temporarily reduced by institutions. Furthermore, foreign institutional investors showed a similar tendency; in September, foreign investors increased holdings of application-side companies like Zhipu AI and MiniMax through Stock Connect, while reducing holdings of computing power companies like Cambricon. Regarding individual investors, retail holdings of AI concept stocks also showed divergence in September. Retail investors tended to favor application-side stocks because their performance is more stable and the risk is lower. Conversely, the proportion of retail holdings in the computing power sector decreased due to high volatility.

IV. Future Outlook: Changes in AI Concept Stock Investment Logic in the Second Half, with the Application Sector as the Core

Looking ahead to the second half of 2026, the investment logic for AI concept stocks will undergo a major change, with the application sector becoming the core trend. The main reasons are as follows:
  • Accelerated Commercialization: By 2026, AI applications will feature richer commercial scenarios, with new growth points emerging in areas such as AI+Education and AI+Retail. The performance of application-side enterprises will gradually be released, supporting stock price increases.
  • Sustained Policy Support: Departments such as the Ministry of Industry and Information Technology and the National Health Commission will continue to issue policies supporting the development of AI applications. For example, in October 2026, the Ministry of Industry and Information Technology will release the "AI Application Scenario List" to encourage enterprises to adopt AI technology to improve efficiency.
  • Technological Breakthroughs: AI large model technology is continuously advancing, with companies like MiniMax and Zhipu AI launching more advanced models to enhance application effectiveness and promote commercialization.
  • Capital Preference: Institutional funds will continue to flow into the application side, driving sector gains. Concurrently, individual investors will follow institutions to increase allocations to the application side.

For investors, the focus in the second half of the year should be on opportunities in the application sector, especially in fields like AI+Healthcare, AI+Finance, and B-side software. Although the computing power sector is under short-term pressure, it still holds long-term investment value and can be positioned during dips.

V. Risk Warning

Despite the broad prospects of AI concept stocks, the following risks must be noted:

  • Policy Risk: Policies in the AI application field may change, such as increased regulation, affecting corporate performance.
  • Technological Risk: AI technology development may not meet expectations, leading to delays in application implementation.
  • Market Risk: AI concept stocks are highly volatile, requiring investors to pay attention to risk control.

In summary, in September 2026, the AI concept stock market will exhibit a differentiated pattern, with the application side becoming a favorite for capital due to accelerated commercialization. In the second half of the year, investors should adjust their investment logic, focusing on opportunities in the application sector while also paying attention to the long-term value of the computing power sector.

Detail Page Ad