Another AI funding round is easy to ignore. CuspAI's new round is harder to file away as routine, because the money is not chasing chat, video or office agents. It is going into materials discovery, one of the slower and harder bottlenecks behind chips, batteries, carbon capture and industrial supply chains.
On July 20, the Cambridge-based startup announced a $450 million Series B at a $2.6 billion valuation and launched the AI Materials Foundry, a global network of data, laboratories, compute and scientific expertise for designing new materials.
The bet is lab speed
CuspAI was founded by Chad Edwards and Max Welling only two years ago. The new round was led by Kleiner Perkins and NEA, with significant participation from Bezos Expeditions. It follows a $100 million-plus Series A less than a year ago.
The company describes itself as a search engine for the material world. A user specifies desired chemical or physical properties, and the system generates candidate materials that still need simulation, synthesis, testing and industrial validation.
“If we don’t make progress fast, the next 50 years of industrial progress will be constrained by a single challenge: the world needs materials that don’t yet exist,” CuspAI's co-founders said.
The Foundry is the point
The AI Materials Foundry brings together more than 45 founding partners. The Guardian and Bloomberg both put the number above 48. The names include NVIDIA, Meta, Samsung, Hyundai Motor Group, Henkel, Applied Materials, Tokyo Electron and Lam Research.
That mix matters. Semiconductor manufacturing, energy grids, batteries, carbon capture and water treatment all depend on materials that can work under cost, performance, safety and manufacturing constraints.
Kleiner Perkins partner Josh Coyne put it simply: most big leaps in technology come down to a material. In AI infrastructure, that is not a slogan. GPUs, HBM, packaging, cooling and power systems all run into materials limits.
The useful read is supply chain
For readers outside the UK, the story is less about a European unicorn and more about the next layer of the AI infrastructure race. The competition has already moved from models to power, cooling, wafers, equipment and supply assurance.
CuspAI is trying to make the discovery loop shorter: generate candidates, test them with partners, and move toward materials that can be built rather than just imagined by a model. That is why government capital, Bezos Expeditions, AMD Ventures and semiconductor-equipment companies can appear in the same story.
The numbers still need discipline. $450 million and $2.6 billion show capital appetite, not industrial deployment. More than 45 Foundry partners provide access and demand, not automatic adoption. The real checks are case studies that move from AI candidate to lab validation, and then into chip, energy or manufacturing tests.
CuspAI's round says part of the next AI race may sit between the laboratory bench and the factory floor, not only on model leaderboards.
Sources: CuspAI announcement, The Guardian, Bloomberg, CocoLoop, NEA investment note; checked the $450 million Series B, $2.6 billion valuation, more than 45 Foundry partners, $100 million-plus Series A, and semiconductor and materials-use claims.