Zhu Xiaohu: The outbreak point of China's AI lies in the application scenario end and requires "going beyond the technology".

Recently, Zhu Xiaohu, the managing partner of Jinsha River Venture Capital, sparked industry discussion for questioning human-shaped Siasun Robot & Automation and AI. He publicly announced plans to withdraw from related projects in bulk, believing that their commercialization path is not clear, their valuation is inflated, and their customers are mostly universities and state-owned enterprises, making it difficult to achieve profitability. In addition, Zhu Xiaohu also put forward the viewpoint that "all AI applications are shell applications," believing that the long-term barrier of AI applications lies not in technology but in activities beyond technology, such as integrating workflows, editing capabilities, proprietary hardware, etc., rather than training the underlying models. A series of viewpoints have triggered wide-ranging discussions in the venture capital circle, with some supporting his warning against the industry bubble, while others criticizing his lack of in-depth understanding of the AI and Siasun Robot & Automation fields. Recently, Zhu Xiaohu once again appeared at the 2025 Shanghai Jiao Tong University Alumni Investment Annual Meeting. Faced with controversy, he said, "I didn't expect that article to have such a wide impact, even my mother forwarded it to me. The heat around artificial intelligence is too high." During the roundtable discussion, he did not mention human-shaped Siasun Robot & Automation, but extensively talked about AI applications in China, saying that the breakthrough point of AI in China lies in application scenarios and requires "moving beyond technology." The explosion of AI application scenarios and opportunities in the Chinese market In today's rapidly evolving AI technology and applications, Zhu Xiaohu believes that the emergence of large models has opened up a huge space for imagination in the industry. From a bottom-up perspective, China's AI development still follows the business strategy of "imitation and micro-innovation." At the current stage, the industry is in the process of scaling from "1 to 100," providing a rich soil for the explosion of application scenarios. Zhu Xiaohu pointed out that China's potential at the application scenario end is particularly significant. Although the scale of domestic investment funds is only 1/4 to 1/5 of that in the United States, the cost-performance advantage is obvious. He cited the example of the U.S. market, mentioning that more than half of AI funding flows into the fields of healthcare, finance, and privatization. There are opportunities in these directions in China as well: for example, in the field of medical image recognition, China has comprehensive and rich medical data to support technological micro-innovations; in education scenarios, applications such as AI-assisted homework correction and personalized learning are emerging; and in the financial advisory field, a service model combining localization requirements also shows potential. Additionally, the maturity of domestic large models has further reduced the development threshold. The high cost problem of relying on overseas technologies like ChatGPT in the past has been alleviated, and developers can now call on locally available models with similar performance at a lower cost. This inclusive technology allows more small and medium-sized teams to participate in innovation, especially in vertical scenarios where rapid trial and error and iteration are important. Zhu Xiaohu emphasized that the key to success on the application side lies in "understanding user needs" rather than simply pursuing technological barriers. Using the example of early mobile internet, he pointed out that many applications initially only imitated existing models, but ultimately successful enterprises were based on a deep understanding of user pain points. Following the strategy and micro-innovation: the unique path of AI development in China Zhu Xiaohu believes that China's achievements in the field of AI cannot be separated from the combination of "follow-up strategy" and "micro-innovation." He admitted that China has not excessively invested in the original field of "0 to 1", but has chosen to optimize and localize on existing technological foundations to reduce trial and error costs and improve commercial efficiency. This strategy has been verified in fields such as electric vehicles, new energy batteries, and chips, and AI fields also follow this logic. For entrepreneurs, he suggests avoiding the core track dominated by large companies and instead looking for differentiated sub-sectors. For example, scenarios like automatically generating meeting minutes or processing legal documents may seem basic but have value due to high-frequency needs. However, these applications have not yet fully landed in China, and entrepreneurs need to design solutions based on localized demand. Taking the example of intelligent assistants (Agents), he pointed out that they are essentially holistic platforms, and future ecological competition will revolve around user access points. Therefore, entrepreneurs need to focus more on commercial logic and social needs beyond technology. Zhu Xiaohu specifically mentioned that the barriers of AI applications are gradually shifting from technology to comprehensive capabilities. Excellent entrepreneurs need to have a background in technology and business sensitivity, and be able to "understand business, understand society, understand human nature." Using the example of grassroots work in law firms, he explained that while AI can replace some processes, the real challenge lies in combining technology with industry rules and meeting compliance requirements. This article is reprinted from the "Finance and Economics Club" public account; GMTEight Editor: Li Fo.