As data centers strain under the heat generated by AI workloads, a new wave of startups is turning to artificial intelligence to design the chips of tomorrow. Discovered Materials, a Y Combinator alum, announced today that it has raised a $9 million seed round led by Lightspeed India Partners, with participation from Peak XV Partners and angel investors including Paul Graham and Gokul Rajaram. The company’s mission: use swarms of AI agents to discover novel semiconductor materials that run cooler, reducing the massive energy and cooling demands of modern data centers.
Founded by Advaith Sridhar and Akash Ramdas, the startup combines Ramdas’s doctorate in materials science from Stanford with Sridhar’s experience building AI agents at Persona AI and Luma Labs. Their software pipeline generates material candidates using Anthropic models in a custom harness, then validates them through simulations run on foundational physics models they’ve trained in-house.
From 20 guesses a day to thousands
Traditional materials discovery is painstakingly slow. During his PhD, Ramdas recalls making perhaps 20 educated guesses per day about promising compounds. With AI agents running around the clock on cloud infrastructure, the company can now explore thousands of research directions daily, guided by the same expert intuition but at machine speed.
“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,” Sridhar told TechCrunch. The company has already released examples of hundreds of new materials and introduced its “Material Discovery Bench,” a benchmark designed to track how frontier AI models handle the challenge of predicting viable substances.
Also read: From red-light district to AI powerhouse: How King's Cross became London's answer to Silicon Valley
Discovering a material is only half the battle. The engineering trade-space is brutal: a compound that dissipates heat well might be impossible to manufacture at scale, or its electrical properties might be compromised. “It’s a bit of playing whack-a-mole with atomic structures,” said Hemant Mohapatra, the Lightspeed partner who led the round. “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.”
Why this matters for the chip industry
The stakes are enormous. Data centers consume roughly 1-2% of global electricity, and a significant portion goes toward cooling the servers that run AI models. Chips that run cooler would not only cut energy bills but also allow for denser computing architectures, potentially extending the life of Moore’s Law in an era of physical limits.
Discovered Materials is entering a crowded field. Companies like MatNex, SandboxAQ, and CuspAI are all pursuing AI-driven materials discovery, with varying focuses. The startup’s bet is that a laser focus on semiconductor thermal problems—rather than a broader approach—will yield commercially viable results faster.
The startup says it has already identified several materials that match the properties of substances currently used by major chipmakers, though it cannot disclose details. When promising candidates emerge, the company plans to patent either the material’s use in GPUs or the manufacturing process, then license those patents to chip manufacturers. Sridhar hopes to have materials worth patenting within the next year.
The long road to commercial impact
Despite the enthusiasm, no AI-discovered drug or material has yet made a significant commercial impact. The closest example is Insilico Medicine’s Renterosib, the first generative-AI-discovered drug to enter Phase II clinical trials. On the materials side, promising candidates like MatNex’s rare-earth-free permanent magnets and new semiconductors from Panasonic and Citrine Informatics have been identified, but none have reached large-scale deployment.
Mohapatra believes the bottleneck isn’t finding candidates—it’s filtering them and synthesizing them in the real world. “A lot of this will involve actually going into wet labs and making things as well,” Sridhar acknowledged. “And this is the process that cannot be sped up.”
For now, Discovered Materials is betting that its combination of expert knowledge, proprietary data, and rapid AI-driven experimentation will give it an edge over well-funded frontier labs. The next year will test whether that bet pays off—and whether AI can finally deliver on its promise of transforming materials science from a slow, manual discipline into a high-throughput engine of innovation.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. The markets and technologies discussed are volatile and subject to rapid change; readers should conduct their own research before making any investment decisions.
