HomeArticle

Former QY Capital investors team up with a professor from the University of Hong Kong to develop an AGI investment machine, securing investment from Monolith | Exclusive from Yingke

华南-彭丽2026-03-23 10:37
With a dual background in finance and AI, the products have been put into practice.

Yingke has learned that Grace Investment Machine (hereinafter referred to as GIM), an AI investment technology company, recently completed an angel round of financing worth tens of millions of yuan, jointly invested by Monolith and Wuyuan Capital. The funds from this round will be mainly used for the technological iteration of the self - developed large financial model, the construction of data and computing power infrastructure, and the continuous expansion of a global top - notch team. Zhiguan Capital will serve as the exclusive financial advisor later.

GIM was founded in July 2025. The company positions itself as a next - generation self - evolving intelligent agent investment platform. It hopes to create an investment machine and an asset management platform in the AGI era through a multi - agent architecture, pioneering a new category of AI asset management. It is one of the first new - generation companies globally to explore Agentic Investing asset management.

GIM's founding team has both investment practice and cutting - edge AI research backgrounds, which gives it unique advantages in this cross - field. The founder, Xu Jiahao, once served as a senior investor at Neumann Advisors and Wuyuan Capital, with over 10 years of investment experience in primary and secondary capital markets. He has led the IPO exits of multiple technology projects. Another founder, Dr. Liu Qi, is currently an assistant professor in the Department of Computer Science at the University of Hong Kong. His research focuses on multi - modal large models, and he is committed to building intelligent systems with complex understanding, reasoning, and interaction capabilities, and promoting the cutting - edge application of artificial intelligence in financial markets. He has worked at leading global research institutions such as DeepMind and Facebook AI Research.

In addition, the core team members of the company mainly come from top technology giants such as Meta, DeepMind, and leading hedge funds like Millennium Management.

Although traditional investment strategies are far more than just quantitative and fundamental analysis, from a research paradigm perspective, the two most representative approaches are: one is quantitative investment, which models and trades based on structured data such as price, trading volume, and various financial factors; the other is fundamental investment, which conducts analysis around company operations and industry trends, combining unstructured information such as research reports, earnings call transcripts, and management communications.

In current practice, AI in the investment field mostly remains in the research - assistance stage. With the development of Agent technology, investment research, strategy generation, and portfolio management are expected to gradually move towards automation and continuous optimization. Xu Jiahao, with more than a decade of investment experience, has noticed this trend: "We believe that an AI - native asset management platform will become a new category. The first generation is subjective investors, such as Warren Buffett; the second generation is quantitative investment, such as Renaissance Technologies; and the third generation will reconstruct the entire investment process through intelligent agents."

Based on this judgment, GIM has been continuously promoting the construction of underlying model capabilities and has launched a self - developed large financial time - series model. This model is trained based on high - frequency market data and models complex market dynamics. It can not only accurately predict future returns in different time windows, extract complex features, and optimize factor combinations but also provide support in trading execution and liquidity capture. As the scale effect and transfer learning ability increase, this capability is evolving from a dedicated model to a more general - purpose investment platform and is expected to further evolve into an automated investment agent covering the entire chain of research, decision - making, and execution under the integration of multi - source data.

"From the very beginning, we have positioned ourselves as a self - evolving intelligent agent investment platform," said Xu Jiahao. This judgment stems from the team's profound accumulation in the traditional quantitative field and the core talent reserve in the direction of multi - modal large models and intelligent agents. Based on the integration of these two capabilities, GIM has built an industrial - level R & D and production system comparable to top - tier investment institutions since its establishment, covering R & D pipelines, organizational structures, and risk governance modules.

While the technical system is gradually taking shape, GIM is also promoting the layout related to private securities investment funds. In the future, the company will successively launch a series of fund strategies and products driven by intelligent agents and attempt to verify the transformation path of AI in the asset management industry with actual performance.

"Financial investment bears real wealth and trust, and we must maintain a sense of awe," Xu Jiahao told Yingke. "GIM will adhere to scientific methods and strict rigor, using AI to deeply integrate the extreme data capabilities of quantitative investment with the logical reasoning of fundamental research to create long - term value for global investors."

Investor's view:

Cao Xi, the founding partner of Monolith, said: "What impressed us most about GIM is that the team has a clear understanding of the AI - native investment platform and is advancing very quickly. The team not only has top - notch AI research capabilities but also truly understands the complexity of finance and asset management. The opportunities brought by large models should not be limited to improving investment research efficiency but have the potential to reconstruct the entire asset management process. We believe they have the opportunity to participate in defining the new paradigm of asset management in the AGI era."