In 2025, the A-share market concluded with an annual gain exceeding 18%, while the Yongying Technology Select Fund, managed by Ren Jie, clinched the title of the year's top-performing public fund with a staggering annual return of 233.26%. This achievement not only marks another instance of a public fund doubling its value within a year but also decisively shattered the championship return rate record, which had been held by Wang Yawei for 18 years. The fund's fourth-quarter 2025 report reveals that it significantly outperformed its benchmark; by the end of the quarter, the net asset values per share for the Yongying Technology Select A and C funds were 3.7795 yuan and 3.7523 yuan, respectively, with quarterly growth rates of 13.18% and 13.01%, while the benchmark returned -2.39%.
However, an examination of the fund's asset allocation in the fourth quarter shows a notable reduction in equity exposure. The third-quarter report indicated that stock holdings constituted 91.59% of the fund's total assets; by the end of the fourth quarter, this proportion had decreased to 78.76%. The remaining assets were primarily held as bank deposits and settlement reserves, and the fund also completely divested its previously small bond holdings.
Specifically, during the fourth quarter of 2025, the top ten holdings of the Yongying Technology Select Fund, in order, were: Shengyi Technology (600183.SH), Zhongji Innolight (300308.SZ), Wus Printed Circuit (002463.SZ), Eoptolink Technology (300502.SZ), Shennan Circuits (002916.SZ), Suzhou Tfc Optical Communication (300394.SZ), Suzhou Dongshan Precision Manufacturing (002384.SZ), Shenzhen Kinwong Electronic (603228.SH), Foxconn Industrial Internet (601138.SH), and Cig Shanghai (603083.SH). Regarding changes in holdings compared to the third quarter of 2025, Suzhou Dongshan Precision Manufacturing, Shenzhen Kinwong Electronic, Cig Shanghai, and Foxconn Industrial Internet entered the top ten list. The fund increased its positions in several stocks, including Wus Printed Circuit, Shennan Circuits, Shengyi Technology, Suzhou Tfc Optical Communication, and Eoptolink Technology. Conversely, T&S Communications (300570.SZ), Montage Technology (688008.SH), Henan Shijia Photons Technology (688313.SH), and Everprox Technologies (300548.SZ) dropped out of the top ten holdings. Additionally, the fund reduced its stake in Zhongji Innolight, with a sequential decrease of 17%.
During the reporting period, the Yongying Technology Select Fund continued to focus its investments on the global cloud computing industry. Building on its third-quarter assessment, which was based on the realization of new technologies by 2027, the fund adjusted its strategy in response to advancements in cutting-edge global AI models and commercialization progress. Manager Ren Jie noted in the report that the launch of a new-generation multimodal model by a leader in the search industry during the fourth quarter led to another significant leap in industry capabilities. With its full-stack layout encompassing chips, cloud computing, models, and applications, this company demonstrated unique differentiated advantages in model iteration efficiency and implementation capabilities, re-establishing its position as a global leader in the AI field.
From the user perspective, Gemini continued to gain market share in the consumer segment. Meanwhile, other model companies maintained healthy development, actively increasing investments to enhance their model capabilities and making continuous progress in both commercialization and capital operations. Looking ahead, Ren Jie believes that models still possess significant potential for improvement in areas such as AI agents, multimodality, and autonomous learning. The application ecosystem is transitioning from scale expansion to structural upgrades. Over the past three years, mainstream model products achieved rapid user growth through swift market penetration. As coverage rates become more saturated, model companies are entering a second phase of development, centered on user stratification, capability differentiation, and service depth.
Specific application directions include high-barrier vertical sectors like healthcare, enterprise and industry-level B2B applications, complex application forms represented by AI agents and multimodality, and deep integration within internet scenarios such as e-commerce and advertising, as well as within the core businesses of leading tech companies. The implementation and scaling of these applications are expected to be key drivers for the improvement of model companies' commercialization metrics. Computing infrastructure is also adapting to the evolving demands of models and applications. Characteristics across different dimensions, such as Mixture-of-Experts (MoE), long-context understanding, multi-agent systems, and the separation of planning and execution, have significantly increased the complexity of data exchange and state management within computing clusters, rapidly elevating the importance of communication and storage overhead within the computational system.
To address constraints related to bandwidth, latency, and energy efficiency, next-generation computing clusters are accelerating the adoption of architectural innovations like Co-Packaged Optics/Optical I/O, Optical Circuit Switches, and orthogonal backplanes. As these new technologies progress from design to large-scale deployment, companies that were deeply involved in the early stages of R&D collaboration and supply chain integration with leading manufacturers are poised to capture more substantial industry benefits during the penetration rate increase. Overall, global AI models are still in a phase of continuous capability enhancement and expanding application scenarios, indicating strong sustainability and certainty for the industry's development. Consequently, the fund will continue to prioritize investments related to the global cloud computing industry, with a particular focus on optical communication and PCB segments within the communications sector.
Comments