A YC Startup Puts $1M Towards 5 Unsolved Materials Science Problems

San Francisco-based Discovered Materials, a startup from Y Combinator’s spring 2026 batch, has unveiled a groundbreaking $1 million challenge aimed at harnessing artificial intelligence to conquer five fundamental, unsolved problems in materials science. The initiative, announced this week, seeks to establish definitive benchmarks for AI's capabilities in understanding and predicting the complex physics governing advanced materials, an area where even the most sophisticated frontier models have yet to make substantial breakthroughs.

The Materials Science AI Frontier

While large-scale AI models have demonstrated remarkable prowess in mathematics and significant progress in the life sciences, the intricate physics of advanced materials has largely remained beyond their grasp. Advaith Sridhar, co-founder of Discovered Materials, highlighted the unique difficulty in defining clear problems within this domain, unlike the precise “yes or no” nature of mathematical challenges. The startup’s goal is to create “goalposts” to measure AI's advancement in this critical field.

To identify these pivotal challenges, Discovered Materials collaborated with an advisory group of prominent academics from leading global universities. This expert panel carefully articulated the five open problems, which collectively represent some of the most stubborn theoretical and practical hurdles in materials science. The company is offering a pooled prize of $1 million for successful solutions.

The Million-Dollar Questions

The five grand challenges that Discovered Materials and its academic advisors have posed are:

• Cuprate Superconductivity: Developing a comprehensive theory to explain cuprate superconductivity, a phenomenon where certain materials exhibit zero electrical resistance at ultra-low temperatures.
• Silicon-Silicon Dioxide Interfaces: Creating a numerical method capable of predicting the exact structures of interfaces between silicon and silicon dioxide, a ubiquitous combination in nearly all modern semiconductors.
• Thermal Properties via Grain Structure: Establishing a numerical method to accurately map a material's grain structure to its thermal properties, which is crucial for designing more effective heat dissipation materials.
• Light-Induced Material Behavior: Formulating a theory to predict why shining light on materials can alter their behavior, particularly in relation to superconductivity.
• Density Functional Theory Improvement: Devising a numerical method to significantly improve density functional theory, a cornerstone technique widely used in materials science for modeling the atomic and molecular structures of substances.

Stanford professor Felipe Jornada, who contributed to the development of these challenges, emphasized their profound potential impact. He noted that solving these observed but unexplained phenomena would not only deepen our fundamental understanding of physics but also directly influence the construction of future electronics, energy storage solutions, and manufacturing processes.

AI's Initial Struggle and the Path Forward

Prior to announcing the challenge publicly, Discovered Materials subjected several frontier AI lab models to these problems. Granting autonomous agents internet access and resources, the company observed that even after days of dedicated effort on each problem, the AI models struggled to achieve significant progress. This initial struggle underscores the complexity of the tasks and highlights the gap between current AI capabilities and the demands of advanced materials science.

Despite the initial setbacks with internal testing, Sridhar expressed that the origin of the solution—whether from Discovered Materials itself, another startup, a university researcher, or a major AI lab—is not the primary concern. The core objective is to advance collective understanding. He explained that these challenges are intended to act as concrete “goalposts,” enabling the scientific community to track whether AI models are truly “climbing in this domain” and fostering breakthroughs that will deepen our grasp of materials and their fundamental behaviors.

This initiative also signals a notable shift in the landscape of scientific inquiry. As AI continues to demonstrate its capacity for uncovering unexpected insights and techniques, the emphasis increasingly falls on posing the right questions. Sridhar articulated this new approach: “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.” The role of the advisory professors was critical in shaping these “right questions.”

Discovered Materials' Broader Mission

Discovered Materials’ primary business revolves around leveraging AI to discover novel materials, specifically for advanced semiconductors. The company collaborates with leading chip manufacturers on research while also pursuing its own scientific endeavors. Sridhar, who had previously indicated hopes of generating a useful material within a year during the company’s seed round announcement in August, now expresses even greater optimism. He stated, “I’m more confident now that we will be able to find these industrially useful materials in a timeline shorter than a year.”

The $1 million materials science challenge represents a strategic move by Discovered Materials to accelerate foundational research that will ultimately benefit its core mission and the broader technological landscape. By incentivizing breakthroughs in these complex areas, the startup is not only pushing the boundaries of AI's scientific application but also laying the groundwork for a new generation of high-performance materials critical for the future of computing and beyond. The outcomes of this challenge could reshape our approach to material discovery, driving innovation across multiple industries.

FAQ

Q: What is the primary goal of Discovered Materials' $1 million challenge? A: The main goal is to incentivize AI models to solve five critical, unsolved problems in materials science, thereby advancing AI's ability to understand complex material physics and setting clear benchmarks for progress in this challenging scientific domain.

Q: Why are materials science problems particularly difficult for AI compared to other scientific fields? A: Materials science problems are often more difficult to specify clearly, lacking the precise “yes or no” answers found in fields like mathematics. They involve complex physics and require a deeper understanding of underlying phenomena, which current frontier AI models have struggled to fully grasp.

Q: How might solving these challenges impact future technology? A: Solutions to these problems could significantly improve our understanding of physics and directly influence the development of next-generation electronics, more efficient energy storage systems, and advanced manufacturing processes, leading to breakthroughs across various industries.