Traditional SEO (Search Engine Optimization) is gradually transitioning to a GEO (Generative Engine Optimization)-driven AI commercialization closed-loop process, causing a shift in traffic entry points and a reshaping of user behavior. This transition is expected to move the application side from the technical validation phase to the commercial value realization phase.
According to Zhitong Finance APP, China Galaxy Securities released a research report stating that the artificial intelligence sector was jointly propelled by three catalysts in January, raising expectations for the overall risk appetite and inflection point. The bank still recommends active positioning in sub-sectors and key companies within the artificial intelligence sector that support earnings. The bank believes that domestic AIDC auctions will start to recover and show an upward trend in Q4 2025. Domestic computing capacity is expected to enter a new cycle due to demand promotion and intensive bidding. Attention should be paid to the segments and enterprises of the domestic computing power industry chain.
The main perspectives of China Galaxy Securities are as follows.
The sector exploded in a big way in January, with AI application tracks surging across the board.
The Artificial Intelligence Sector Index (884201.WI) rose 10.86% in January, compared with 3.76% for the Shanghai Composite Index, 1.65% for the CSI300 Index, and 4.47% for the ChiNext Index. In January, the transaction volume of the artificial intelligence sector was RMB 2795.823 billion, an increase of 152.67% year-on-year and an increase of 94.74% month-on-month. The bank believes that the artificial intelligence sector experienced a strong recovery in January, with trading volumes continuing to increase. Subsequently, the sector showed structural differentiation and entered a phase of consolidation. January was mainly influenced by three catalytic factors: industrial, policy and capitalization developments.
1) Industry: CES 2026, launch of NVIDIA’s Rubin architecture for implementing AI in the physical world, and major collaboration between NVIDIA and Eli Lilly to develop fundamental models for AI-assisted drug discovery. 2) Policy aspect: Eight departments announced “Special Action Implementation Opinions on Artificial Intelligence + Manufacturing.” 3) Capitalization progress: Zhipu AI and MiniMax were listed on the Hong Kong Stock Exchange one after another. In January, the artificial intelligence sector was influenced by these three catalysts, which jointly increased risk appetite and inflection point expectations across the sector. The bank still recommends active positioning in sub-sectors and key companies within the artificial intelligence sector that support earnings.
From SEO to GEO, the paradigm shift in search in the AI era is accelerating commercialization at the application end.
ByteDance’s large model, DouBao, will have more than 100 million DAUs by the end of 2025, making it the largest AI application in the country by user base. AI applications are moving from technological novelty to large-scale practical application, gradually changing the way people access information and make decisions. From traditional web search SEO to GEO, AI applications are driving fundamental changes in marketing and traffic distribution logic. The bank believes that there is a significant increase in focus on AI applications as the application side of AI receives high-density catalysis. Traditional SEO (Search Engine Optimization) is gradually transitioning to GEO (Generative Engine Optimization), which is expected to drive a closed-loop process of AI commercialization, triggering the migration of traffic entry and reshaping of user behavior, and driving the application end from the technical validation phase to the commercial value realization phase.
Domestic computing capacity is expected to enter a new cycle due to demand promotion and intensive bidding.
NVIDIA’s H200 chips are expected to enter the Chinese market under certain conditions, which will be beneficial to the development of domestic computing chips and ecosystem in the long run. The bank believes that domestic AIDC auctions will start to recover in the fourth quarter of 2025 and show an upward trend. Domestic internet giants will accelerate data center deployment in 2026, potentially even earlier than 2025. Once the H200 supply resumes, it will improve the efficiency of large-scale model training, further accelerate the implementation of AI applications, and increase the demand for domestic computing chips in inference. Domestic computing capacity is expected to enter a new cycle.
investment advice
We will focus on the following subtracks and companies: 1. Domestic computing power industry chain. 2. IDC Service Provider and Computing Power Lease. Domestic IT innovation manufacturer, 3. 4. AI agents and applications. 5. Cloud Computing Vendors. 6. All-in-one machine and edge AI. 7. Data element industry chain enterprises involved in supply, distribution and application. 8. Eda.
risk warning
Risk that technology iterations fall short of expectations. risks of increased competition among tech giants; Legal and Regulatory Risks. Supply chain risk. The risk that downstream demand will fall short of expectations.
