Discovered Materials Announces $1 Million AI Challenge for Five Unsolved Materials Science Problems

San Francisco-based Discovered Materials, a Y Combinator spring 2026 startup, has unveiled a $1 million prize pool for AI models that can solve five long-standing problems in materials science.

The initiative targets gaps that have limited the impact of frontier AI on the physics of advanced materials, a field that underpins semiconductor manufacturing, energy storage and emerging quantum technologies.

What Is the $1 Million AI Challenge?

Discovered Materials gathered an advisory panel of professors from leading universities to define five open research questions that have resisted conventional approaches. The company will award the total prize money to any team that delivers a verifiable solution, whether the work comes from an academic lab, a corporate research group or an independent AI developer.

List of the Five Problems

• A theory to explain cuprate superconductivity, the phenomenon where certain copper-oxide materials conduct electricity with zero resistance at very low temperatures.
• A numerical method to predict the atomic structures that form at interfaces between silicon and silicon dioxide, a junction present in virtually every semiconductor device.
• A numerical method to map grain-scale structure to thermal conductivity, enabling the design of materials that dissipate heat more efficiently.
• A theory to predict how illumination changes material behavior, especially in contexts that affect superconductivity.
• A numerical method to improve Density Functional Theory, the computational framework used to model the electronic structure of atoms and molecules.

Why Focus on Materials Science Now?

Frontier AI models have already solved celebrated mathematics puzzles and contributed to drug discovery, yet they have not yet demonstrated comparable breakthroughs in the complex, multiscale physics that governs advanced materials. As co-founder Advaith Sridhar explained, "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."

Materials scientists frequently encounter phenomena that are observable but lack a predictive theory. Stanford professor Felipe Jornada, who helped craft the challenge list, said, "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."

How the Challenge Was Tested

Before publicizing the prize, Discovered Materials released its own frontier lab models to work on the five questions. The AI agents were given internet access and autonomous compute for several days on each problem, but they made little progress. The experiment demonstrated that current models still need significant advances or larger compute budgets to tackle these deep-physics questions.

"It's not particularly a concern to us whether we find it, or someone else finds it, or a professor finds it, or the frontier labs find it," Sridhar said. "The reason we're putting out these challenges is because we want them to serve as the goalposts. If you go and solve these challenges, we are deepening our understanding of materials and how things around us behave."

Implications for the Semiconductor Industry

Discovered Materials' core business is to discover new substances that can be used in next-generation chips. The company already collaborates with major chip manufacturers on research projects and plans to fund its own scientific work. Solving any of the listed problems could shorten the design cycle for new semiconductor materials, improve heat management in high-performance processors, and enable novel device architectures that rely on superconductivity or photo-responsive behavior.

When the startup announced its seed round in August, Sridhar told TechCrunch that the goal was to produce a useful material within a year. "Those timelines have accelerated a little bit," he says today. "I'm more confident now that we will be able to find these industrially useful materials in a timeline shorter than a year."

What Researchers and AI Teams Need to Know

Participants must submit a solution that can be independently validated. For theoretical problems, this means a peer-reviewable proof or a model that reproduces known experimental data. For numerical methods, the submission should include open-source code, benchmark results and a clear description of the algorithm's advantages over existing techniques.

The prize pool will be distributed based on the difficulty and impact of each solution, with the advisory board overseeing the evaluation. No single deadline has been announced; the challenge will remain open until all five problems are resolved or the pool is exhausted.

What to Watch for Next

Discovered Materials will track progress through a public leaderboard that records when a team claims a breakthrough. The company expects that larger AI models from organizations such as OpenAI or Anthropic could make rapid advances if directed toward these challenges.

Stakeholders, including chip designers, energy-storage firms and academic labs, should monitor announcements from the challenge to gauge when AI-generated insights become reliable enough for commercial adoption.