China Galaxy Securities has released a research report stating that with intensive catalysts in the AI application sector, attention on AI applications has significantly increased. Traditional SEO is gradually transitioning to GEO-driven AI commercialization closed loops, which is expected to propel the application sector from a technology validation phase to a period of commercial value realization. The logic for AI application commercialization is anticipated to first see scale in B2B applications. The rationale for GEO leading the AI application trend lies in brands competing for control over traffic in the AI era, thereby enhancing the commercial value of GEO. Simultaneously, the cost-reduction and efficiency-improvement effects of applying AI models within enterprises are gradually becoming apparent. It is recommended to focus closely on the explosion of AI applications in the B2B sector.
With intensive catalysts in the AI application sector, GEO is reshaping the traffic logic of the AI era. Leading Hong Kong-listed large model companies MiniMax and Zhipu AI have shown strong performance post-listing. Nvidia and pharmaceutical giant Eli Lilly have formed a collaboration potentially worth up to $10 billion over the next five years, establishing a lab aimed at advancing the development of foundational models for AI-assisted drug discovery. OpenAI launched ChatGPT Health and announced the acquisition of healthcare startup Torch. China Galaxy Securities believes that alongside these intensive catalysts, AI application focus has significantly heightened. The gradual shift from traditional SEO to GEO-driven AI commercialization closed loops is expected to push the application end from technology validation to commercial value realization.
B2B AI applications are expected to lead the initial surge, while long-term value should be sought in B2C AI application targets. The report posits that the commercialization logic for AI applications will likely achieve scale first in B2B applications. GEO's leadership in the AI application trend stems from the battle among brands for traffic control in the AI era, which in turn boosts GEO's commercial value. Meanwhile, the cost-saving and efficiency gains from deploying AI models internally within companies are increasingly visible. It is advised to prioritize focus on the breakout of AI applications in the B2B sector, such as AI+Marketing, AI+Industrial Software, AI+Healthcare, and AI+Finance. Furthermore, traditional high-quality B2C product application companies possess user bases and brand influence; empowering their products with AI is expected to further solidify their moats. It is recommended to pay attention to B2C application companies with long-term investment value.
Data center demand release and intensive tendering suggest domestic computing power may enter a new cycle. Domestic AIDC tendering began to recover and show an upward trend in Q4 2025. In 2026, major domestic internet companies are expected to accelerate their data center deployment plans, potentially at a faster pace than in 2025. If this is accompanied by restored supply of H200, it would enhance large model training efficiency, further accelerating the implementation of AI applications and the demand for domestic computing power chips in the inference phase. Domestic computing power is poised to enter a new cycle.
The report recommends focusing on the following targets: 1) Large model and MaaS providers: Alibaba-W, Zhipu AI, MINIMAX-WP, iFlytek, SenseTime, TALKWEB, Kunlun Tech. 2) Domestic computing power and data center industry chain: Cambricon, Hygon, Loongson, Runze Tech, Dawei Tech, Sinnet, Baosight Software, ISoftStone, Inspur, Digital China, Envicool, KSTAR, Zhongheng Electric. 3) AI+Marketing: BlueFocus, OneNet One创, Visual China, Mobvista. 4) AI+Industrial Software: Digiwin Intelligent, Nancal Technology, Supcon, Empyrean, Semitronic, Primarius. 5) AI+Healthcare: XtalPi, Insilico, Jiahui, Winning Health. 6) AI+Office: Comet, Kingsoft Office, Wondershare, Foxit. 7) AI+ERP: Kingdee, Yonyou. 8) AI+Finance: Hundsun, Tonghuashun, Compass, Fortune.
Risk warnings include the risk of AI application iteration falling short of expectations, intensifying industry competition, legal and regulatory risks, supply chain risks, and risks of downstream demand underperforming expectations.
Comments