CuspAI Raises $450 Million and Launches AI Materials Foundry

The financing and foundry are related parts of the same expansion. CuspAI describes the foundry as a network of more than 45 industrial companies, research laboratories, and data and compute providers. Named participants include NVIDIA, Meta, Samsung, Hyundai Motor Group, Applied Materials, Tokyo Electron, Lam Research, universities, and public research organizations.

CuspAI's platform, MIRA, is designed to help partners specify desired material properties, generate candidate structures, run simulations, plan synthesis routes, and route candidates toward experimental validation. The company says NVIDIA will contribute accelerated-computing infrastructure and Meta's research team will contribute atomistic modeling technology. These are announced roles and partnerships, not evidence that every proposed material will translate into a manufacturable product.

CuspAI's own announcement confirms the round size, valuation, lead investors, and geographic expansion plans. The Guardian independently reported the financing and the participation of Jeff Bezos and the UK government's Sovereign AI Fund. Tech Funding News also reported the $450 million round, the $2.6 billion valuation, and the foundry launch.

The public material also includes a company case study with Finnish chemicals group Kemira. CuspAI says it screened a very large molecular search space and delivered 20 candidates for further testing in six months. That example is useful context, but it remains a company-reported result rather than an independently reproduced benchmark.

The round shows investors backing AI systems that connect foundation models and scientific data to physical experimentation, not only text or code generation. For teams evaluating similar platforms, the hard questions sit after candidate generation: data provenance, simulation accuracy, synthesis feasibility, laboratory throughput, protection of partner data, and the rate at which model suggestions survive experimental validation.

CuspAI now has substantial capital and a broad partner network, but the important operational measure will be repeatable delivery of validated materials across multiple industries. Procurement and research teams should separate the scale of the funding and coalition, which are documented, from claimed reductions in discovery time, which should be checked against project-specific baselines and independent results.