初创公司Discovered Materials近日完成了900万美元的种子轮融资，投资方包括Lightspeed India Partners和Peak XV Partners，天使投资人阵容包括Paul Graham、Gokul Rajaram和Thariq Shihipar[1]。该公司由材料科学博士Akash Ramdas和AI专家Advaith Sridhar联合创办[1]，后者曾在Persona AI和Luma Labs从事Agent相关工作[1]。
该公司采用AI代理群体的方法来发现新材料，以提高芯片效率并降低热量[1]。Discovered Materials使用Anthropic模型生成材料候选，并通过自训练的物理模型对其进行验证[1]。这套方法大幅提升了研发效率——从Ramdas博士期间每天约20次猜测，提升到现在每天数千次[1]。公司已发布了数百种新材料示例和材料发现基准工具[1]。
在商业化方向上，Discovered Materials计划对GPU中的材料使用或芯片制造工艺进行专利保护和许可授权[1]，并希望在未来一年内拥有可专利的新材料[1]。
Discovered Materials, a materials science startup, has raised $9 million in seed funding from Lightspeed India Partners to develop next-generation materials that improve chip efficiency and reduce heat generation [1]. The company, co-founded by Akash Ramdas, a materials science PhD from Stanford, and Advaith Sridhar, an AI specialist with experience at Persona AI and Luma Labs, employs AI agents to accelerate the discovery process [1].
The startup's approach represents a significant leap in research velocity. The team leverages Anthropic models to generate material candidates, which are then validated through self-trained physics models [1]. This workflow has dramatically increased the pace of discovery—from approximately 20 candidate evaluations per day during Ramdas's doctoral research to thousands daily [1]. The company has already published hundreds of new material examples and released a materials discovery benchmark tool [1].
Discovered Materials plans to monetize its findings through patent protection and licensing agreements tied to material applications in GPUs and semiconductor manufacturing processes [1]. The startup aims to secure patentable new materials within the coming year [1]. The seed round also included backing from Peak XV Partners and angel investors including Paul Graham, Gokul Rajaram, and Thariq Shihipar [1].