Y Combinator startup offers $1 million for solutions to five materials science problems

=== HEADLINE ===

Y Combinator startup offers $1 million for solutions to five materials science problems

Five open problems

In mathematics, problems often have clear formulations and verifiable answers. In materials science, those benchmarks are harder to define: researchers observe physical phenomena they cannot yet explain. To choose the problems, Discovered Materials assembled a group of scientific advisers from leading universities around the world.

Prizes for solutions come from a total pool of $1 million. The contest covers five areas:

Stanford University professor Felipe Hounada, who helped develop the problems, said that explaining observed phenomena would improve our understanding of physics. The answers could also influence the development of electronics, energy storage and the production of new materials.

Problems as benchmarks for AI

Before announcing the contest, Discovered Materials tested the problems on leading models. AI agents were given access to the internet and other resources and allowed to work autonomously for several days on each problem. They made no substantial progress. Startup co-founder Advait Shridhar said OpenAI or Anthropic might be able to try solving the problems in a single run if they gave the models significantly more computing resources or used a newer model.

For the company, however, it does not matter who finds a solution: a contest participant, a professor or a leading lab. Shridhar said the problems should serve as benchmarks for tracking whether models are improving in this area. Solving them should also deepen understanding of materials and the physical phenomena around us.

The contest reflects a shift in how scientific research is approached: first, problems and ways to test solutions are defined; then researchers look for explanations. Shridhar stressed that framing the right question in materials science is difficult. Professors who helped prepare the problems made that possible for the startup.

The startup’s goal is new materials for chips

Discovered Materials’ primary focus is developing new materials for semiconductors. The startup works with major chipmakers on research projects while also funding its own scientific work.

When the company announced its seed round in August, Shridhar told TechCrunch he expected to produce a useful material within a year. He now says the timeline has accelerated: he is more confident that the company can find materials with industrial applications in less than a year.

Dataist's take: The $1 million prize pool gives AI concrete scientific benchmarks, but initial tests showed that current models have yet to make notable progress toward solutions. For the startup, the contest complements its work on semiconductor materials. But the timeline for producing materials with industrial applications remains its estimate, not a stated guarantee.