The landscape of investment management is rapidly evolving, driven largely by the integration of artificial intelligence (AI) technologies. One of the key players in this transformation is DeepSeek, a platform that has garnered significant traction not only among institutional investors but also within the realm of public funds. This breakthrough signifies a larger trend where AI is reshaping various aspects of investment strategies and decision-making processes.

In recent months, public fund organizations have increasingly embraced the potential benefits of AI in their investment research and decision-making practices. Reports indicate that many companies have begun implementing a variety of open-source models, with DeepSeek being a prominent example. This transition is not just about enhancing the technical capabilities of fund managers; it also provides retail investors with innovative tools to explore new investment avenues.

The central question for investors, however, remains whether the implementation of AI will translate into improved performance for mutual funds. In a survey conducted by the Securities Times, it was revealed that fund managers—especially those focused on quantitative strategies—are particularly enthusiastic about AI's potential to deliver alpha, or excess returns, surpassing market benchmarks.

According to the findings, quantitative fund managers have emerged as key advocates of AI's incorporation into investment strategies. These managers, responsible for overseeing both actively managed quantitative funds and enhanced index funds, have been actively investigating ways to leverage AI for superior market performance. Several fund managers shared insights into their early explorations of AI, highlighting their commitment to exceeding market averages through advanced computational techniques.

Examining specific fund performance illustrates the tangible advantages of integrating AI systems. For instance, the Huitianfu Guozheng 2000 Enhanced Index A fund, Anxin Quantitative Selected CSI 300 Enhanced Index fund, and Haifutong CSI 300 Enhanced Index fund have all achieved exceptional returns, placing them at the forefront of their respective categories. Over the past year, the Anxin and Haifutong funds ranked as the top performers within the best-performing CSI 300 Enhanced Index funds, boasting excess returns of 15.55% and 8.77%, respectively.

Anxin's fund manager, Shi Rongsheng, underscores the importance of utilizing machine learning and deep learning methodologies, which he fully adopted upon his appointment in August 2023. His objective is to capitalize on pricing inefficiencies in the market, thereby enhancing its overall efficiency through the statistical advantages offered by AI.

Haifutong's strategy is based on an AI-driven stock selection model, employing a sophisticated approach that mixes deep learning with multi-factor scoring systems to effectively capture nonlinear market information. Their satellite strategies, which involve event-driven and sector rotation tactics, further bolster the fund's adaptability and performance.

The Huitianfu Guozheng 2000 Enhanced Index A fund demonstrates remarkable results, achieving a positive return of 58.83% over the past year and an excess return of 28.5%, which is the highest among its peers. According to its fund manager, the strategy operates by combining multi-factor models with AI-driven stock selection methodologies, streamlining the evaluation process for individual stocks.

On the front of actively managed quantitative funds, several other funds, including Bosera's ESG Quantitative Stock Selection and CMB's Quantitative Selected funds, have also reached historical highs in their net values. The revitalized activity within the A-share market has created a fertile ground for quantitative investing, with AI increasingly taking center stage in this domain.

Over the past six months, the rapid advancements in AI technologies have prompted fund managers to deeply acknowledge the value of these tools. Many quantitative fund managers have adopted an increasingly proactive stance toward assimilating AI into their investment research processes. Lin Lihe, the manager of the Haifutong CSI 300 Enhanced Index fund, highlighted the efficiency and holistic analytical advantages that AI-driven models provide in a fast-changing market. The new models are designed to self-learn and adapt based on evolving market conditions, enabling them to respond swiftly to emerging trends and information.

The increasing popularity of DeepSeek has only reinforced fund managers' resolve to explore AI applications within investment strategies. Shi Rongsheng expressed confidence that large language models and similar technologies will significantly extend the boundaries of quantitative investing. Following the implementation of cost-reduction strategies for API usage in August 2024, he began leveraging DeepSeek's APIs to enhance functionality across various areas, including programming efficiency, accelerating machine learning deployments, and uncovering alternative factors for investment.

Nonetheless, practical challenges remain in applying AI to investment strategies. The nuances of quantitative investing complexity make it an intricate endeavor. Issues such as low signal-to-noise ratios and non-stationarity in data can complicate matters, with many aspects—ranging from algorithm selection and hyperparameter tuning to variable preprocessing—affecting the efficacy of AI applications in quantitative investing.

The ongoing exploration of AI in finance has kept public fund firms at the forefront of innovation. As the technologies behind models like DeepSeek reach maturity, many firms are undertaking substantial revisions of their investment research processes. At the beginning of 2024, Bosera Funds identified DeepSeek's capabilities for automatic coding and logical reasoning after extensive research. They became the first to implement DeepSeek-V1 on their self-operated combustor servers, later upgrading to DeepSeek-V2 in August 2024. With the anticipated arrival of DeepSeek-R1 in early 2025, Bosera is poised to explore its applications across investment research, advisory services, and software development.

Furthermore, Yingying Fund reported that DeepSeek plays a critical role in optimizing essential business operations within the fund industry. Through its ability to rapidly analyze vast amounts of research reports and extract vital information, DeepSeek has significantly enhanced research efficiency, enabling analysts to focus more on critical decision-making tasks.

The implications of AI transcend mere operational efficiencies; they may fundamentally reshape the investment frameworks within which fund managers operate. Currently, most quantitative investment strategies in public funds remain entrenched in traditional linear multi-factor models. Some self-styled "quantitative" strategies are overly reliant on fund managers' subjective decisions, often referred to as "human quantification."

However, as talent with relevant educational backgrounds increasingly enters the public fund space and breakthroughs in AI technology continue to emerge, quantitative investment teams are gaining substantial strength. A growing cohort of fund managers is striving to upgrade their investment frameworks, seeking to mitigate emotional biases and volatility's adverse impacts on portfolio management.

For instance, the fund manager of the Huitianfu Guozheng 2000 Enhanced Index fund underscored that, in the investment management process, it is essential to minimize direct intervention in portfolio holdings. The aim is to translate investment ideas into structured, rule-based quantitative strategies and cautiously adjust those strategies to influence portfolio allocations indirectly. This approach is intended to safeguard the replicability of investment management while avoiding negative emotional influences on investment outcomes.