YC Startup Launches $1M AI Challenge to Crack Five Hard Materials Science Problems

San Francisco–based Discovered Materials, a member of Y Combinator's spring 2026 cohort, has unveiled a $1 million prize pool for AI systems that can solve five long-standing problems in materials science. The challenges were curated with input from academic advisors at top global universities and target phenomena that researchers can observe but not yet explain.

The five problems

• A theory for cuprate superconductivity — zero electrical resistance at ultra-low temperatures.
• A numerical method to predict silicon/silicon-dioxide interface structures, ubiquitous in semiconductor devices.
• A method linking grain structure to thermal properties for better heat-dissipation materials.
• A theory explaining how light exposure alters material behavior, especially superconductivity.
• An improved numerical approach for Density Functional Theory, the workhorse of atomic-scale modeling.

Frontier models from leading labs were given days of autonomous compute on each problem and made limited progress. Discovered Materials says the purpose is not to claim the solutions but to establish durable benchmarks — goalposts that reveal whether AI is truly climbing the materials-science capability ladder.

Why it matters for GPU / AI infrastructure: Breakthroughs in any of these areas directly affect the hardware stack AiGpu customers rely on. Better interface models accelerate transistor scaling; grain-to-thermal maps enable cooler, denser GPU clusters; superconductivity insights could one day slash interconnect losses; and a more accurate DFT speeds the discovery of novel dielectrics and channel materials for future AI accelerators.