Discovered Materials puts $1M behind five unsolved materials problems

Discovered Materials, part of Y Combinator's spring 2026 batch, is trying to turn materials science into a benchmark. Unlike math, where problems have precise definitions and yes-or-no answers, materials science lacks an equivalent of the Millennium Prize Problems, as co-founder Advaith Sridhar noted. So the startup assembled academics to define five challenges, then tested frontier models on them first — giving AI agents internet access and other resources to work autonomously for days. The models did not make much headway.

That result sits behind the prize structure. Sridhar said it does not matter to the company whether it, a professor, or the frontier labs find the solutions. The point is to create goalposts that show whether models are climbing in this domain, and to shift scientific inquiry toward verification and question-setting first, with ideas to explain the problem coming later.

The company's main business is developing new materials for semiconductors, partnering with chip giants while funding its own research. When it announced its seed round in August, Sridhar hoped to generate a useful material within a year. He now says those timelines have accelerated and he is more confident the company will find industrially useful materials in less than a year.