As AI chips push data centers to their thermal limits, a new startup is using artificial intelligence to help design the next generation of cooler semiconductors. Discovered Materials is building an AI-driven discovery pipeline aimed at finding materials that can improve heat management in integrated circuits.
The company recently raised a $9 million seed round led by Lightspeed India Partners, following its time in Y Combinator. Peak XV Partners and several angel investors also joined the round, signaling growing interest in AI-powered materials research.
Founded by Advaith Sridhar and Akash Ramdas, the startup combines large language models with physics-based simulation tools. Its system generates candidate materials, then tests them through models trained to estimate whether those substances could work in real chip environments.
According to the founders, the approach can produce thousands of material ideas each day, far beyond the pace of traditional research workflows. The company has also introduced a benchmark called Material Discovery Bench to evaluate how frontier AI models perform in this field.
Discovered Materials says its focus is not only on discovering new substances, but also on finding candidates that can be manufactured and integrated into chips at scale. That balance between performance, stability, and production remains one of the most important challenges in semiconductor innovation.
If successful, the startup plans to pursue patents and license promising materials to chipmakers. The broader goal is to help shape a future where AI does not just power advanced computing, but also helps build the materials that make it more efficient.
This kind of research could accelerate a new era of smarter, cooler, and more sustainable computing systems.
