CuspAI, a UK-based AI materials-discovery startup, raised $400 million at a roughly $2.6 billion valuation in a round backed by Jeff Bezos' Bezos Expeditions and Kleiner Perkins. The company applies generative AI to design novel materials on demand -- starting with sorbents for carbon capture and extending to batteries, catalysts, and semiconductors -- compressing discovery timelines that traditionally span years of lab trial-and-error.
The thesis is that materials science is where AI's search-and-simulation strengths pay off most directly: the space of possible compounds is astronomically large, and the cost of physically testing each is high. A model that can propose candidates likely to have target properties turns an intractable search into a tractable one. That is exactly the kind of hard problem that justifies a deep-tech moat.
“Bezos and Kleiner co-leading at a $2.6 billion mark on an early-stage company says the bar for conviction is high, but the ceiling is higher.”
The round is a marker for where smart capital is rotating: away from undifferentiated AI application wrappers and toward AI-for-science plays with defensible technical depth. Bezos and Kleiner co-leading at a $2.6 billion mark on an early-stage company says the bar for conviction is high, but the ceiling is higher.