Since the launch of ChatGPT by OpenAI, the tech industry has witnessed an unprecedented surge in enthusiasm for AI models, particularly generative modelsMajor tech corporations like Microsoft and Google have heavily invested in this domain, with stocks of companies such as Microsoft, NVIDIA, and TSMC experiencing significant increases, effectively spearheading a notable bull market in U.SequitiesThe ripple effects of this enthusiasm have not gone unnoticed globally, with Chinese tech giants like Tencent, Alibaba, and Baidu eagerly stepping into the arena, determined to carve out their respective futures in AI.

However, while the competition for AI supremacy predominantly involves tech firms in both the domestic and international spheres, one notable exception has emerged: the Chinese quantitative firm, Huanshan Quantitative Trading, which stands behind the domestic AI model DeepSeek

This involvement of a quant firm in the AI model space has caught the attention of many in the investment community.

Huanshan Quantitative, a leader in the quant sector, has had its share of criticism within the A-share marketThe firm has earned accolades for its status in China’s quantitative trading landscape, with area-specific saying: "North is Jiukun, and South is Huanshan," pointing to their leading rolesHuanshan Quant has made remarkable achievements, having been recognized as the first private quantitative firm in China to surpass the 100 billion yuan mark in assets under management and remains the only one to break through this significant threshold to date.

Yet, Huanshan Quantitative is not merely a traditional quantitative firm; it has also diversified its portfolio into the realm of AI modelsThe company established Huanshan AI back in 2019, demonstrating an early commitment to exploring the intersections of AI and quantitative strategies

• Through Rate Cuts, Capital Flees: U.S. Stock Market Faces Volatility
• Nasdaq Hits 20,000: A New Benchmark?
• Silicon Valley Bank Collapse Sparks Financial Crisis Fears
• Polysilicon Futures Launch Upcoming
• U.S. Stock Futures Stabilize

By 2020, Huanshan AI revealed that they had invested close to 200 million yuan into their self-developed deep learning training platform known as "Firefly One," equipped with 1,100 GPUsFollowing this, the investment ramped up to 1 billion yuan for "Firefly Two," featuring around 10,000 NVIDIA A100 GPUs, positioning them within China’s elite computing rankings.

Then, towards the beginning of 2023, Huanshan Quantitative initiated an independent organization specifically focused on developing large models, named DeepSeekThis division aims to pursue genuinely human-level artificial intelligenceTheir ambition goes beyond simply replicating existing models like ChatGPT; it includes probing the uncharted territories of AGI (Artificial General Intelligence).

In May 2024, DeepSeek announced the open-sourcing of its second-generation Mixture of Experts model, DeepSeek-V2. This model impressively claimed to rival the performance of GPT-4 Turbo while being priced at a mere 1% of GPT-4, earning it the moniker of "Price Butcher." Moreover, a recent update concerning DeepSeek includes the unveiling of DeepSeek V2.5-1210, which is expected to represent the final update to the V2 series before they embark on a new foundational model series.

DeepSeek asserted that this latest update employs post-training techniques to enhance the model's capabilities across various areas, such as mathematics, coding, writing, and role-playing scenarios

It also incorporates optimized file upload functionalities and supports online search capabilities, significantly bolstering its applicability across diverse work and life contexts.

Investment circles are captivated by the notion of a quantitative giant venturing into AI models, prompting speculation about the feasibility of “AI-driven stock trading.” In response to such assertions, Huanshan Quantitative's CEO, Lu Zhengzhe, clarified in 2023 that their AGI exploration is not intended for trading purposes but is centered on developing GTP-related large models, distinctly separating their technological pursuits from financial activitiesThis sentiment was echoed by founder Liang Wenfeng, who emphasized that their foray into large AI models is fundamentally unconnected to their quantitative and financial interests.

Nevertheless, the evolution of the AI industry has prompted several private equity firms to explore AI’s utilization in investment landscapes

For instance, as of late 2017, virtually all of Huanshan's quantitative strategies incorporated AI models into their calculations, indicating that AI has played a pivotal supportive role in their investment decisions.

As of now, most products under Huanshan Quantitative have displayed positive returns, with a few showcasing particularly remarkable profitsAdditionally, other private equity firms such as Zhi Yu Zhi Shan Investment announced in June 2023 their intention to task four researchers and an AI-based robot, tentatively named Cybertron, with independently managing five distinct private equity fundsThe Cybertron fund is overseen by the company’s general manager, showcasing a growing inclination towards AI in fund management.

On top of that, data from private equity sites indicate that Zhi Yu Zhi Shan’s CTA1 fund has achieved an impressive return rate of 276.07% this year, further underscoring the potential of AI-driven investment models

However, it is essential to note that while AI robots have begun taking on independent roles in fund management, their current function remains largely supportive, typically centering around stock selectionThis approach subtly differs from traditional quantitative methods, wherein fund firms provide relevant factors to tech companies like Microsoft, allowing AI to select suitable stocks, often in a “black box” manner where exact stock specifics remain opaque.

The race to dominate AI in the quant sector has become a widely accepted goal among industry playersFor instance, Jiukun Investment launched its "Wutong Program," targeting global graduates with exceptional academic backgrounds to select the most promising new quant talentJiukun is also recognized for its early and comprehensive incorporation of AI technologies throughout its quant investment processes.

Similarly, Heiyi Asset has been at the forefront of integrating AI strategies since 2017, establishing its dedicated algorithmic AI team