Workflow
Agentic AI(智能体人工智能)系统级操作平台
icon
Search documents
MetaNovas获A+、A++两轮融资,以智能体「军团」加速新材料开发 | 36氪首发
36氪· 2026-03-09 00:11
Core Insights - The article discusses how MetaNovas is revolutionizing the development of new materials through the use of AI, moving away from the traditional "blind screening era" [2][8] - The company has successfully completed multiple funding rounds, indicating strong investor confidence in its innovative approach [3] Group 1: Company Overview - MetaNovas has recently completed A+ and A++ funding rounds with investments from notable firms such as Hillhouse Capital and Kangaroo Mom Group [3] - The company utilizes an Agentic AI system to enhance efficiency in the development of consumer-oriented new materials, significantly reducing the time required for product development from 3-5 years to as little as 12 months [4][8] Group 2: Technology and Innovation - The AI-driven platform developed by MetaNovas allows for multi-objective decision-making, taking into account performance, process requirements, and regulatory constraints [4] - The company has created a molecular language generation model that can efficiently generate molecules across a vast chemical space, achieving over 95% efficiency in molecular generation [4][5] Group 3: Data and Learning - High-quality data accumulation and an active learning system are crucial for improving model accuracy, with training data sourced from literature, academic collaborations, and internal experiments [5] - The internal experimental platform not only provides successful validation data but also includes "failure" data, which enhances the AI system's precision during iterations [5] Group 4: Market Strategy and Applications - MetaNovas' AI system has improved the success rate of molecular recommendations to over 60%, significantly lowering trial-and-error costs [8] - The company is expanding its development pipeline to include bioactive ingredients, medical materials, functional polymers, and more, while collaborating with CDMOs for production scaling [8]