Discovered Materials to use AI agents to find better materials for integrated circuits

### AI Startup Discovered Materials Raises $9 Million to Develop Cooler, More Efficient Chips

AI startup Discovered Materials is using artificial intelligence agents to search for new materials that could help make integrated circuits more efficient and reduce the heat generated by advanced AI chips. The company believes new semiconductor materials could help address one of the growing challenges facing AI data centers: rising power consumption and the need for increasingly sophisticated cooling systems. Discovered Materials recently raised $9 million in a seed funding round led by Lightspeed India Partners after emerging from Y Combinator. The round also included investment from Peak XV Partners and prominent angel investors Paul Graham, Gokul Rajaram and Thariq Shihipar. The startup plans to use the new funding to expand its team, strengthen its laboratory capabilities and scale its artificial intelligence research agents. Founded by Advaith Sridhar and Akash Ramdas, Discovered Materials operates cloud-based autonomous AI agents capable of exploring thousands of virtual material possibilities every day. The system uses custom model frameworks that incorporate advanced AI models to generate potential new materials. Those candidates are then tested through physics simulations to evaluate properties such as stability, dielectric constants and thermal performance. The founders bring complementary experience to the company, with Ramdas holding a doctorate in materials science from Stanford and Sridhar having previously worked on AI agents at Persona AI and Luma Labs. The startup says its automated approach dramatically increases the speed at which researchers can explore potential materials. Instead of manually testing a limited number of ideas each day, its AI agents can operate continuously in the cloud and investigate thousands of possibilities. Discovered Materials has also introduced the Material Discovery Bench, a platform designed to evaluate how advanced AI systems perform against real-world semiconductor material challenges. The company intends to identify promising candidates, develop intellectual property around them and eventually license the resulting technologies to major chipmakers. Its research is particularly focused on solving thermal challenges associated with modern semiconductor devices. As AI workloads become more demanding, powerful processors generate enormous amounts of heat, increasing the electricity required for cooling and creating additional infrastructure costs for data centers. Discovered Materials believes better materials could help improve heat management while maintaining the electrical performance required by modern chips. The startup says it has already identified several materials that appear to match properties of materials currently used by major semiconductor manufacturers, although it has not disclosed specific details. However, discovering a promising material is only the first step toward commercial use. A substance that performs well in simulations may prove difficult or expensive to manufacture at semiconductor scale. It could also introduce unwanted electrical characteristics that make it unsuitable for real-world chip production. This creates a complex optimization problem in which thermal, electrical and manufacturing requirements must all work together. Lightspeed partner Hemant Mohapatra described the process as a difficult search across atomic structures, where a material becomes valuable only when multiple performance requirements converge. Discovered Materials is competing in an emerging field that includes companies such as MatNex, SandboxAQ and CuspAI. However, the startup says its focus on semiconductor thermal management differentiates it from other material-discovery platforms. As AI models become increasingly capable of predicting new substances, competition in the field is expected to intensify. The company plans to seek patents covering promising materials and their applications in GPUs or manufacturing processes. It would then license those technologies to semiconductor manufacturers around the world. Discovered Materials hopes to identify new materials suitable for patent protection within the next year. If successful, its technology could provide chipmakers with new ways to manage heat, improve energy efficiency and support the growing computing demands of artificial intelligence.