Discovered Materials secures seed funding to accelerate semiconductor discovery
The startup uses high-frequency simulation and custom physics models to identify materials capable of reducing heat in high-performance integrated circuits.
Julian Reeve
Aug 7, 2026 · 1 min read
A doctorate in materials science once allowed for roughly 20 hypotheses a day. Discovered Materials, a new entrant in the semiconductor materials space, is now running thousands of simulations daily. The company recently closed a $9 million seed round led by Lightspeed India Partners to expand a software pipeline that replaces human guesswork with automated inquiry.
The technical architecture relies on large language models to generate material leads, which are then processed through proprietary foundational physics models. These simulations verify whether a candidate material possesses the thermal properties required for next-generation AI workloads. As data centers face increasing power and cooling requirements, the ability to engineer more efficient integrated circuits at the molecular level has become a primary bottleneck.
Founders Advaith Sridhar and Akash Ramdas, drawing on experience from Stanford and Persona AI, are positioning the firm to solve the heat problem that current hardware scaling has created. By operating agents 24/7 on the cloud, the startup aims to shorten discovery cycles for materials that were previously too complex or time-consuming to model manually.