Radical AI uses autonomous labs to accelerate alloy discovery

As the United States seeks to decrease reliance on China for critical minerals, New York-based company Radical AI is leveraging artificial intelligence to develop innovative alloys that minimize dependence on supply-constrained metals. The firm has created an autonomous laboratory platform employing AI and robotics to design, manufacture, and test advanced materials, dramatically accelerating the discovery process for alloys in aerospace, defense, and industrial applications.

In September, Radical AI joined forces with Sandia National Laboratories to investigate metal-hydrogen interactions and uncover new materials for hydrogen separation from waste gas streams. Operating from its Brooklyn Navy Yard facility, the company has showcased a rapid alloy discovery process, generating 300 new compositions within roughly 16 weeks, according to CEO Joseph Frause.

This work, supported by RTX and Nvidia, could provide US manufacturers with an alternative means to tackle supply chain vulnerabilities by redesigning materials based on more readily available elements.

Radical AI's strategy differs from Washington's focus on expanding domestic mining, processing, and refining capacity to mitigate dependence on foreign suppliers. The company operates downstream, utilizing refined metals to create advanced materials tailored to specific industrial performance requirements. Rather than engaging in mining or mineral processing, the technology identifies element combinations that deliver desired properties while reducing exposure to supply-constrained metals.

For instance, hafnium, a key element in high-temperature aerospace and defense alloys, is largely sourced from China, posing a supply risk. Radical AI has successfully developed high-performance alloys with minimal hafnium content, addressing this vulnerability.

By employing AI-driven material design in conjunction with physical testing, the company aims to significantly shorten the traditional 10-15 year mining process, which can cost over $100 million for a single material system. The autonomous experiment generation, sample synthesis, and property analysis cycle allows for rapid iteration, with machine learning continuously improving alloy designs based on experimental results.

Ultimately, Radical AI strives to reduce materials development timelines from decades to just a few years, enabling manufacturers to develop application-specific alloys that address critical mineral availability and performance requirements from the outset.