Radical AI is a deep-tech company focused on transforming materials science and advanced manufacturing through AI, with a mission to accelerate the discovery, development and production of new materials. Materials R&D has historically moved slowly because each candidate must be synthesized and physically tested, a process that can take years. Radical AI's approach is to compress that loop by combining AI-driven prediction with autonomous robotic experimentation in a tight, self-improving cycle.
The company's system uses AI to screen vast spaces of possible materials, potentially billions of candidates, to identify the most promising options. It then synthesizes and tests those candidates in a fully autonomous lab capable of running an estimated 50 to 100 times more experiments than traditional setups. Critically, the experimental data captured from each run is fed back into the models, creating a closed-loop, self-driving lab that improves its predictions over time, the kind of compounding data advantage that distinguishes leading AI-for-science efforts.
Radical AI has emphasized building real physical infrastructure, establishing what New York State officials described as the state's first fully autonomous materials-science labs at the Brooklyn Navy Yard, an effort highlighted by the governor's office. This combination of frontier AI with operational autonomous labs positions the company to target applications across energy, advanced manufacturing and decarbonization, where novel materials are a key bottleneck.
In July 2025 Radical AI announced a $55 million seed (described as a Series Seed+) led by RTX Ventures, with participation from NVentures (NVIDIA's venture arm), Noa, Infinite Capital, Eni Next and AlleyCorp, and reporting indicates a subsequent Series A. The company is led by CEO Joseph F. Krause alongside co-founder Jorge Colindres. Radical AI competes in the autonomous-materials-discovery space alongside players such as Periodic Labs and CuspAI, differentiating through its emphasis on operating real autonomous labs that close the loop between AI prediction and physical experimentation.