Partnering with Radical AI to Advance Materials Science with Interpretability

We're excited to announce a new partnership between Radical AI and Goodfire to fundamentally dismantle the black box of AI-driven materials discovery and design.

Radical AI integrates proprietary AI modeling with in-house autonomous laboratory systems to create a powerful self-learning, closed-loop system for materials discovery, testing, and development. They recently announced a $55 million Series Seed+ funding round led by RTX Ventures with participation from NVentures (NVIDIA's VC arm) and other investors.

Radical AI and Goodfire are building generative models to intelligently forge materials for a specific purpose, directly creating a material based on its desired function. AI has already transformed materials science by enabling faster molecular modeling, property prediction, and materials screening—delivering simulated predictions in minutes that would take traditional methods like density functional theory (DFT) days or weeks to compute. Our work now approaches the even more ambitious goal of inverse materials design.

Current materials science AI models remain largely opaque, limiting researchers' ability to understand why certain predictions work, how models arrive at their conclusions, or what fundamental principles they've learned about structure-property relationships. This gap represents a rich opportunity to pioneer a new sector in materials science for both scientific insight and model improvement.

Our partnership combines Radical AI's materials expertise in AI-modeling with Goodfire's cutting-edge interpretability expertise and platform infrastructure to change this paradigm. By applying interpretability methods to materials models, we extract the underlying physical and chemical principles that these systems have learned, leading to new scientific insights about materials behavior and enabling more controllable, explainable AI-driven materials discovery.

"Applying interpretability to scientific models is one of our core focuses, and is a fundamentally new way of driving scientific discoveries. We're deeply excited to expand that work to the domain of materials science with Radical AI." —Eric Ho, CEO, Goodfire

"For Radical AI, when developing new materials that will enable the new industries of the future, it is critical to have the ability to deeply understand the reasoning behind the AI engine making recommendations of structures to send to the lab." —Joseph Krause, CEO, Radical AI

This collaboration represents a significant step towards AI that is not only predictive but truly interpretable, bridging the gap between correlation and causation. We believe this accelerates not just materials discovery, but our fundamental understanding of how AI systems learn to reason about the physical world.

More details about specific research directions and outcomes will be shared as the partnership progresses.