Radical AI raises $55M seed led by RTX Ventures to build AI-powered materials science R&D facility
Jul 18, 2025 with Joseph Krause
Key Points
- Radical AI closes $55 million seed round led by RTX Ventures to build an autonomous materials R&D facility capable of running hundreds of thousands of experiments per year.
- Co-founder Kristin Persson brings a proven template from her fully autonomous robotic lab at Lawrence Berkeley National Lab, which runs 55 experiments daily and established the Materials Project.
- The investor syndicate of defense prime RTX, Nvidia, and financial VCs reflects a bet that AI-accelerated materials science will unlock R&D breakthroughs across semiconductors, aerospace, energy, and defense.
Summary
Radical AI closed a $55 million seed round led by RTX Ventures, the corporate venture arm of Raytheon Technologies. Nvidia's venture arm, Lux Capital, Infinite Capital, and AlleyCorp also participated. AlleyCorp led the pre-seed and returned for this round. The company ranks this as the third-largest equity-only seed round in New York City history.
Radical was co-founded by a materials scientist trained at Rice University and the Army Research Lab, who grew frustrated watching research stall between academia and commercial scale. He cold-emailed Kevin Ryan at AlleyCorp with a pitch on materials science as the most important untapped AI application. The company launched from AlleyCorp's incubator in March 2024 with a $10 million pre-seed.
The third co-founder is Kristin Persson, a Berkeley professor with an H-index of 182 who helped establish the Materials Project under the Materials Genome Initiative. She built a fully autonomous robotic lab at Lawrence Berkeley National Lab capable of running 55 experiments per day. That lab is the academic template Radical is scaling.
Use of funds
The $55 million funds hiring across AI, materials science, and automation, plus construction of what the company describes as the most advanced materials R&D facility in the world. The target is hundreds of thousands of experiments per year across multiple material systems. The facility runs active learning loops, with experimental data feeding back into the AI engine in real time to direct the next round of experiments.
AI for materials has been bottlenecked by missing training data. Radical is building the proprietary dataset itself through the robotic facility, which functions as both product and competitive moat.
The investor composition reflects the thesis clearly. RTX and Nvidia are buying exposure to a platform that could accelerate R&D across semiconductors, aerospace, defense, energy, and automotive. Materials science underlies all of those verticals.