British startup CuspAI secured $450 million at a $2.6 billion valuation to advance AI materials discovery, drawing major support from Jeff Bezos and Nvidia. The funding round highlights a growing push to apply artificial intelligence to physical world challenges.
Monday’s announcement confirms Kleiner Perkins and NEA led the massive financing, with significant participation from Bezos Expeditions. This Jeff Bezos investment signals strong confidence in using machine learning to solve hardware bottlenecks. Additional backers include Glade Brook Capital Partners, Lux Capital, AMD Ventures, and Britain’s Sovereign AI Venture Fund. The capital will fuel the startup’s mission to simulate novel materials, a process that could drastically reduce the time required to find breakthroughs in semiconductors and clean energy.
Nvidia Joins AI Materials Discovery Push
Central to the startup’s growth is the launch of its “AI Materials Foundry,” a comprehensive network integrating data, laboratories, compute power, and scientific expertise. The company relies on its proprietary AI platform to simulate how novel materials perform, narrowing down vast arrays of theoretical possibilities for physical lab testing. This approach bypasses the traditional, time-consuming trial-and-error phase of materials science.
A major Nvidia partnership underpins this foundry initiative. Nvidia is among the 45 organizations providing the compute infrastructure necessary to run these complex simulations. Meta’s Fundamental AI Research Team is also joining the initiative, underscoring the tech industry’s recognition that software advancements are increasingly bottlenecked by physical hardware limits.
“As AI transforms the physical world, new materials will open up new frontiers across semiconductors, energy and advanced manufacturing,” said Chad Edwards, CEO and cofounder of CuspAI. “The AI Materials Foundry brings NVIDIA accelerated computing infrastructure together with world-class chemistry and materials expertise to help power the next generation of materials discovery.”
Josh Coyne, a partner at Kleiner Perkins, emphasized that major technological leaps rely on physical materials. “Most big leaps in technology come down to a material, and the next set, from cheaper carbon capture to semiconductors and cleaner water, is stuck waiting on materials nobody has discovered yet,” Coyne said in a statement. “CuspAI built a search engine that changes that.”
The Rise of AI in Physical Engineering
The startup represents a broader industry shift where AI models target physical engineering rather than just digital content or software workflows. Jeff Bezos is heavily invested in this specific intersection through his own company, Prometheus. Founded in 2025, Prometheus has raised $12 billion at a $41 billion valuation to create AI specifically for invention and physical engineering. The alignment between Bezos Expeditions and this new venture suggests a strategic focus on tangible, physics-based AI applications.
To support its ambitious goals, the company is expanding its global footprint. The Britain-based startup is opening a new office in Singapore while growing its teams across the U.K., the Netherlands, Germany, Japan, and the U.S. This global expansion aims to connect computational research with localized manufacturing hubs and academic talent.
Accelerating Semiconductor Materials Development
The immediate impact of this funding and the foundry will likely manifest in the semiconductor sector. As traditional Moore’s Law scaling slows, chipmakers are desperate for new semiconductor materials to improve efficiency, manage heat, and reduce reliance on scarce resources. AI materials discovery offers a pathway to identify these compounds faster than traditional research allows.
The involvement of AMD Ventures is particularly noteworthy. As a major chip designer, AMD has a direct interest in finding new materials that could enhance processor performance or lower production costs. The Nvidia partnership also bridges the gap between compute providers and the physical hardware that powers their data centers.
What Happens Next
Over the next 12 to 18 months, watch for the foundry’s first major material discoveries to move from simulation to physical lab testing. If the AI models can successfully predict material behaviors for carbon capture or cleaner water filtration, it will validate AI materials discovery as a viable alternative to decades-long traditional research and development cycles.
Furthermore, the collaboration between Nvidia and Meta signals that infrastructure providers are preparing for a future where AI workloads are heavily focused on scientific simulation rather than just language processing. The success of this Jeff Bezos investment could trigger a wave of capital into startups focused on the physical sciences, fundamentally changing how hardware is built.
— David Kim, technology desk, AXO News