AI workloads generate excessive heat, forcing data centres to burn through electricity and rely on heavy cooling systems. Discovered Materials is now using swarms of AI agents to hunt for new materials that could build more efficient integrated circuits.
The startup raised $9 million in a seed round from Lightspeed India Partners. It emerged from Y Combinator with backing from Peak XV Partners and angels Paul Graham, Gokul Rajaram, and Thariq Shihipar.
Founders Advaith Sridhar and Akash Ramdas launched the firm. Ramdas holds a doctorate in materials science from Stanford. Sridhar previously worked on agents at Persona AI and Luma Labs.
They built a software pipeline using Anthropic models in a custom harness to generate material leads. Foundational physics models trained by the team then run simulations to verify if the candidates are actually useful.
“[Ramdas] was doing maybe 20 guesses a day during his PhD,” Sridhar told TechCrunch. “We’re able to do thousands of guesses a day now by having these agents run 24/7 on the cloud, exploring research directions that he gives them.”
Discovered Materials released examples of hundreds of new materials today. They also launched their “Material Discovery Bench” to track how frontier models handle the challenge.
Companies like MatNex, SandboxAQ, and CuspAI have launched similar efforts. Discovered Materials bets that focusing specifically on thermal problems in semiconductor materials is the path to success. The startup says it has already discovered several materials matching the properties of existing ones used by major chipmakers but cannot share more details.
One challenge is the engineering trade-space. A material that reduces heat generation might be too difficult to manufacture or compromise electrical properties.
“It’s a bit of playing whack-a-mole with atomic structures,” Hemant Mohapatra, the Lightspeed partner who led this round, told TechCrunch. “A material is only useful in the real world if all of them converge at once, which is what makes this a really interesting search problem.”
Mohapatra expects predicting novel substances will become commoditized as models improve. The difference with Discovered Materials is Ramdas’ deep experience in the field and the ability to run a lab that can rapidly experiment and validate candidates. The founders have already done this with several new materials.
When they find valuable candidates, Sridhar says the company will attempt to patent the use of the materials in GPUs or the process by which chips can be made out of the substance. He hopes to have new materials worth patenting in the next year.
However, we still have not seen drugs or materials discovered by AI make a commercial impact. The closest is perhaps Insilico Medicine’s Renterosib, the first drug discovered with generative AI to enter a Phase II clinical trial. Promising candidates have been found on the materials side, like MatNex’s rare-earth free permanent magnets or new semiconductor materials worked out by Panasonic and Citrine Informatics. These have not been commercially deployed at scale yet.
These techniques may be coming into their own as AI continues to improve. This is one reason why Mohapatra says he does not believe finding more candidates is the hold-up for AI materials science. Instead, “filtering them correctly and synthesizing them is the bottleneck.”
While Sridhar believes Discovered Materials’ unique data and expertise will help the startup compete with deep-pocketed frontier labs, he acknowledged that a lot of this will involve actually going into wet labs and making things. This is the process that cannot be sped up.
