YC Startup Offers $1 Million to Crack Five Unsolved Materials Science Problems With AI
YC startup Discovered Materials offers $1 million for AI solutions to five open materials problems spanning chips, superconductors and thermal design.
Summary
On November 5, 2026, months-old San Francisco startup Discovered Materials, from Y Combinator’s spring 2026 batch, announced a $1 million prize pool for solving five open materials science problems. An academic advisory group including Stanford professor Felipe Jornada defined the targets: explain zero-resistance cuprate superconductivity at ultra-low temperatures; predict interface structures between silicon and silicon dioxide, a pairing used in almost all semiconductors; map grain structure to thermal properties for improved heat dissipation; explain light-induced material changes, particularly around superconductivity; and improve Density Functional Theory for modeling atoms and molecules. Jornada said solutions could reshape electronics, energy storage and materials manufacturing.
Frontier AI agents received internet access and other resources, then worked autonomously for several days on each challenge without much progress. OpenAI or Anthropic could attempt them with newer models or more computing power, although neither has announced plans. Co-founder Advaith Sridhar said the problems provide measurable goalposts in a field whose questions lack mathematics-style yes-or-no definitions, supporting a verification-first approach in which researchers establish tests before seeking explanations. Discovered Materials develops semiconductor materials, partners with major chip companies and funds internal research. After announcing its seed round in August, Sridhar targeted an industrially useful material within a year and now expects to reach that milestone sooner.
Positives
- $1 million in prizes gives AI developers and researchers a concrete incentive to tackle five foundational materials science problems.
- Five academically defined challenges create measurable benchmarks for AI progress in a field that lacks an equivalent to the Millennium Prize Problems.
- Solutions could improve semiconductor interfaces, chip cooling, electronics, energy storage and future materials manufacturing.
- Discovered Materials partners with major chip companies while funding its own semiconductor materials research.
- Advaith Sridhar now expects an industrially useful material sooner than his original target of one year after the August seed round.
Risks & concerns
- Frontier AI agents made little progress despite receiving internet access, other resources and several autonomous working days per problem.
- Materials science problems resist precise yes-or-no definitions, making both challenge design and solution verification harder than in mathematics.
- Cuprate superconductivity and light-induced changes around superconductivity remain observed phenomena without adequate theoretical explanations.
- OpenAI and Anthropic have not announced attempts, leaving the impact of newer models or greater computing power untested.
- Discovered Materials has accelerated its useful-material target, but no successful industrial material was announced with the challenge.