"ZhiYuan ShenLan" Secures Angel Round Funding to Build a Data-Driven AI Biomolecular Design Platform | Exclusive Report by 36Kr

36Kr learned that "ZhiYuan ShenLan" recently completed an angel-round financing of tens of millions of yuan. The lead investor was Woyan Capital, followed by Tianfeng Capital and other angel investors. Existing shareholders Inno Angel Fund and Lingyi Venture Capital continued to participate in the investment. The funds will be mainly used for the construction of a generative AI platform for biomolecules and an autonomously driven molecular function evolution platform, as well as for business market expansion.

ZhiYuan ShenLan was founded in 2024, incubated by MegaGenius Technology. It focuses on data-driven biomolecule design and manufacturing. Its founder, Dr. Wang Chengzhi, once served as the chief scientist of MegaGenius Technology and has been working in the life sciences field for more than two decades.

Currently, generative AI (GenAI) is triggering a profound transformation in the life sciences field. Wang Chengzhi once told 36Kr that a "qualitative leap" in this field is approaching. As AI technology evolves from an auxiliary tool to an autonomous platform, the R & D paradigm of the entire life sciences is shifting from "large-scale trial and error" to "precise design and creation." He believes that to truly achieve rational design in life sciences, engineering and digitalization must be realized first.

In the few years after the emergence of AlphaFold 2, it has predicted the structures of more than 200 million proteins, covering almost most proteins of all known living organisms on Earth, with a high degree of credibility. However, in the process of practical application, the core concern of the industry is the function of proteins, not just the intermediate protein structures.

ZhiYuan ShenLan chose "function" as the object of optimization and explored the functional requirements in actual application scenarios. Currently, high-precision protein structure data is relatively scarce. Therefore, it has built an autonomously driven (self-driving) automated experimental platform to efficiently generate functional data. Its biomolecule design system combines the automated experimental platform with AI algorithms, enabling AI to iterate rapidly with real functional feedback and improving R & D efficiency.

Wang Chengzhi told 36Kr that in the future, teams that can quickly and massively produce high-quality biological experiment data are more likely to have high-performance AI models. In the past, automated and high-throughput experiments were mainly for improving efficiency. But in the AI era, they can not only quickly provide experimental results for R & D personnel but also efficiently generate structured and iterable data, directly serving model training and optimization.

Based on this judgment, ZhiYuan ShenLan is trying to build a data-driven bioengineering and molecular design platform to promote the evolution of AI for Science from the 2.0 "navigation-style design engine" to the 3.0 "scientific intelligent autonomous platform."

In the 2.0 era of AI for Science, with the increase in data volume and the progress of algorithms such as deep learning, through the verification and iteration based on the "closed-loop of wet and dry experiments," the accuracy of AI in specific tasks has been significantly improved.

In the future 3.0 era, AI will be able to autonomously design, execute, and iterate the entire scientific research experiment closed-loop. Human scientists will mainly be responsible for tasks at the front and back ends, that is, proposing key questions, formulating strategic directions, and controlling risks. When this stage arrives, the R & D in life sciences will, like APP development in the Internet era and intelligent agent development in the AI era, move towards technological equality and democratization.

Specifically, this autonomously intelligent evolution platform requires three key breakthroughs.

"Unified coordinate system: Translate computing, data, experimental instruments, and production equipment into a form understandable by AI. The iteration of theoretical models is synchronized with the experimental progress, receiving experimental results in real-time and automatically updating parameters. Autonomous decision-making AI Agent: As the scientific research brain, it disassembles complex problems and automatically designs experimental schemes to verify the minimum hypothesis. Automated intelligent experimental platform: As the hands and eyes of AI, it provides a physical basis for large-scale and highly reliable research."

Based on the combination of these three, ZhiYuan ShenLan has proposed a "ten-step" roadmap for AI4S 3.0: from learning existing human knowledge, proposing hypotheses, and verifying experiments to finally making scientific discoveries beyond human intuition in multiple scientific fields.

In the field of biomolecule generation and prediction, through generative AI, researchers can identify new targets, optimize molecular structure design, simplify the pre-clinical verification process, accelerate R & D in many fields such as drug R & D and new material molecular design, and improve the efficiency and innovation level of the entire industrial chain.