Discovered Materials raises $9M to run Anthropic-model agents 24/7 hunting cooler chip materials — thousands of guesses a day versus 20

TechCrunch·medium signalXBlueskyLinkedInCopy link

Y Combinator alumni Advaith Sridhar and Akash Ramdas closed a $9 million seed led by Lightspeed India Partners with Peak XV Partners and angels including Paul Graham, Gokul Rajaram and Thariq Shihipar. The architecture is worth noting for anyone building agent systems: Anthropic language models inside a custom framework generate candidate materials, and custom physics-based simulation models act as the verifier — an LLM proposer plus a deterministic checker, not an LLM judge. Sridhar frames the gain in throughput terms: '[Ramdas] was doing maybe 20 guesses a day during his PhD. We're able to do thousands of guesses a day now by having these agents run 24/7 on the cloud.' The company has released hundreds of materials and launched a 'Material Discovery Bench' tracking frontier-model performance.