CuspAI raises $450M at $2.6B valuation for AI materials discovery, enlists Nvidia, Meta, Samsung in foundry coalition
The AMW Read
CuspAI is a new entrant in AI-driven materials discovery; the $450M round and industry coalition meaningfully advance the segment, with capital and compute implications.
CuspAI raises $450M at $2.6B valuation for AI materials discovery, enlists Nvidia, Meta, Samsung in foundry coalition
CuspAI, a two-year-old Cambridge-based startup using AI to accelerate materials discovery, has raised $450 million in Series B funding at a $2.6 billion valuation — a fivefold jump from its $520 million valuation last September. The round was co-led by Kleiner Perkins and NEA, with participation from Bezos Expeditions, AMD's venture unit, and the UK's Sovereign AI Venture Fund. The company also unveiled its AI Materials Foundry, a coalition of more than 48 companies including Nvidia, Meta, Samsung, Hyundai, Applied Materials, Tokyo Electron, and Lam Research, pooling compute and scientific resources to develop new materials for semiconductor manufacturing.
Why it matters: This funding event signals that AI for physical science — materials discovery, chemistry, advanced manufacturing — is entering a capital-intensive phase distinct from pure software AI. CuspAI's rapid valuation growth from $520M to $2.6B in under a year mirrors the capital-compression arc seen in foundation-model labs, but applied to an industrial bottleneck: the semiconductor industry's reliance on rare metals like iridium and ruthenium. The coalition structure, with Nvidia contributing compute infrastructure and Meta providing atomic simulation models, extends the hyperscaler-distribution moat pattern beyond software into physical R&D. The involvement of the UK sovereign fund also places this within the broader sovereign-AI push in Europe.
Grounded expert take: CuspAI's trajectory exemplifies a recurring pattern where AI-native scientific-discovery companies attract both strategic corporate partners and mega-round venture capital, compressing timelines typically associated with deep-tech hardware startups. The 48-member foundry coalition — spanning chipmakers, equipment manufacturers, and automotive OEMs — gives CuspAI a distribution and validation advantage that is rare for a two-year-old company. However, the $2.6B valuation will hinge on whether the MIRA platform can repeatedly deliver validated candidates at industrial scale, not just in one engagement with Kemira. The hiring of John Giannandrea and advisory roles for Geoffrey Hinton and Yann LeCun signal that the company is positioning itself as a frontier AI lab for physical sciences, not merely a software tool.
