CuspAI

CuspAI’s commercial approach is organized around its “AI Materials Foundry,” a network that combines compute, data, and lab capabilities (including NVIDIA-provided compute infrastructure and Meta Fundamental AI Research Team work on Universal Model for Atoms, UMA) so partners can run end-to-end discovery cycles inside private Foundry instances with data confidentiality. CuspAI targets industrial R&D organizations and public-sector R&D labs working on materials bottlenecks in areas such as semiconductors and advanced electronics, clean energy and energy storage, carbon capture and climate technologies, and advanced manufacturing materials. In its public materials, CuspAI highlights “synthesis-aware” generative modeling paired with physics-based simulation and closed-loop learning from validated experimental outcomes. A key differentiator claimed by the company is data advantage: CuspAI states it has exclusive AI training rights to foundational experimental materials datasets (including the Cambridge Structural Database via CCDC and the Inorganic Crystal Structure Database via FIZ Karlsruhe), and it states it licenses access to materials science content from Wiley and other publishers, then compounds that with validated experimental results returned through Foundry programmes. As of mid-2026, CuspAI has completed a large Series B financing and publicly launched the AI Materials Foundry network with 45+ founding organizations, signaling a strategy to industrialize AI-driven discovery by making the full workflow—generation, simulation, synthesis planning, and experimental validation—available as an integrated “foundry” across regions (the company describes hubs in the United States, Europe, and APAC).