Discovered Materials Launches $1M AI Challenge for Science

A San Francisco startup backed by Y Combinator is offering $1 million in prizes for AI systems that can crack five fundamental problems in materials science. Discovered Materials, which graduated from YC's spring 2026 batch, announced the challenge on November 5, 2026, targeting a domain where frontier AI models have made limited progress despite breakthroughs in mathematics and life sciences.

Why Materials Science Resists AI Solutions

While large language models have solved complex mathematical proofs and advanced drug discovery, materials science presents unique challenges that current AI systems struggle to address. The field lacks the clear problem definitions found in mathematics, where solutions yield definitive yes-or-no answers.

"It's more difficult to specify problems clearly in materials science because it's not like math, where you have a very specific definition and can get to a yes or no," said Advaith Sridhar, co-founder of Discovered Materials. The startup recruited academic advisors from leading global universities to formulate five open problems that represent critical gaps in understanding how materials behave at the quantum and atomic levels.

Officially Launched on November 5, 2026

The challenge went live with a total prize pool of $1 million available for solutions to any of the five problems. Before the public announcement, Discovered Materials tested frontier lab models on the challenges, granting AI agents internet access and computational resources to work autonomously for multiple days. None of the current generation models made meaningful headway on any problem.

"Across materials science, there are phenomena we observe but cannot explain," said Felipe Jornada, a Stanford professor who helped develop the challenges. "Explaining these phenomena would improve our understanding of physics, and the answers would directly shape how we build electronics, store energy, and manufacture the materials of the future."

Benchmark for AI Progress in Physics

The challenge serves a dual purpose: incentivizing breakthrough solutions while establishing clear goalposts for measuring AI advancement in materials science. Unlike the Millennium Prize Problems in mathematics, the field has lacked standardized benchmarks for evaluating computational approaches.

Sridhar emphasized that the competition's value extends beyond who solves the problems first. "The reason we're putting out these challenges is because we want them to serve as the goalposts," he explained. "If you go and solve these challenges, we are deepening our understanding of materials and how things around us behave."

The startup expects major AI labs like OpenAI and Anthropic may attempt to solve the problems using significantly more compute power or next-generation models. Tracking these efforts will reveal whether AI capabilities are genuinely climbing in the materials domain.

Semiconductor Materials Discovery Focus

Discovered Materials focuses on developing new substances for semiconductor manufacturing, partnering with chip industry giants on research projects while funding independent scientific work. The company's core business involves using AI models to identify materials with properties useful for advanced chip production.

When the startup announced its seed funding in August 2026, Sridhar projected the company would discover an industrially useful material within one year. By November, those timelines had accelerated. "I'm more confident now that we will be able to find these industrially useful materials in a timeline shorter than a year," Sridhar said.

Reframing Scientific Inquiry

The challenge reflects how AI is changing fundamental approaches to scientific research. Rather than human researchers formulating both questions and answers, the new paradigm involves experts defining verification criteria and challenges first, then allowing AI systems to generate explanatory theories.

"This is a new way of thinking about science, where we come up with the verification and the challenges first, and then we let somebody come up with the ideas that explain the problem," Sridhar said. He noted that identifying the right questions to ask in materials science requires deep domain expertise, which is where academic advisors proved essential.

What This Means

The $1 million materials science challenge establishes the first standardized benchmark for evaluating AI progress in a domain critical to semiconductor manufacturing and energy storage. With frontier models currently unable to make headway after days of autonomous work, the problems highlight real limitations in AI reasoning about complex physical systems. Whether solutions come from academic researchers, AI labs, or hybrid approaches, the challenge positions materials science as a key test case for next-generation AI capabilities that extend beyond pattern recognition into fundamental physics.