British AI startup CuspAI has secured $450 million in Series B funding at a $2.6 billion valuation as it accelerates efforts to use artificial intelligence for advanced materials discovery in the semiconductor and industrial sectors.

The Cambridge-based company also launched the AI Materials Foundry, a coalition of more than 48 technology companies, industrial manufacturers, and research organizations, including Nvidia, Meta, and Hyundai Motor Group. The initiative aims to combine computing infrastructure, scientific expertise, and laboratory capabilities to develop software that can help researchers discover new materials faster and at lower cost than traditional methods.

CuspAI’s latest funding round was led by Kleiner Perkins and NEA, with significant participation from Bezos Expeditions, the investment fund of Amazon founder Jeff Bezos. The company’s valuation has increased fivefold from approximately $520 million in September 2025, reflecting growing investor interest in AI-driven scientific research.

The startup plans to use a large portion of the capital to expand laboratory operations with Foundry partners in Cambridge, Singapore, and the San Francisco Bay Area. It is also increasing investment in materials discovery, with a major focus on identifying alternatives to rare metals such as ruthenium and iridium that currently pose supply-chain risks for chip manufacturing.

CuspAI originally focused on materials for carbon capture and water purification. However, the company shifted its strategy over the past year after receiving strong demand from semiconductor manufacturers seeking new materials for next-generation AI hardware.

“Chipmakers and semiconductor suppliers are frantically searching for new materials,” said Chad Edwards, co-founder and chief executive of CuspAI. “We’ve been literally pulled by all four limbs.”

Nvidia, which joined the Foundry coalition, expects to collaborate with CuspAI and other industrial partners on new materials for applications ranging from data-center cooling to energy storage systems. Geetika Gupta, Senior Director at Nvidia, said companies are increasingly treating materials research as a strategic competitive advantage.

The company is attempting to differentiate itself through its scientific and AI expertise. Co-founder Max Welling is a prominent AI researcher, while AI pioneers Yann LeCun and Geoffrey Hinton serve on CuspAI’s advisory board. The startup has also appointed semiconductor veteran Abhi Talwalkar, a board director at AMD, to its board of directors and hired former Google and Apple executive John Giannandrea to help establish its California operations.

Despite the strong investor backing, CuspAI acknowledges that commercial breakthroughs remain a work in progress. One of its earlier projects involved building a library of 300 trillion potential metal-organic framework structures for carbon capture applications. The system narrowed the candidates to a small set of promising molecules, but laboratory testing did not yet produce materials that outperformed existing commercial solutions.

The company is now applying the same AI-driven approach to water purification and the removal of persistent “forever chemicals”, with industrial partner Kemira expected to synthesize and test new candidate materials during 2026.

Industry experts view AI as a powerful accelerator for materials research, although they caution that laboratory validation remains a major bottleneck. CuspAI’s strategy of combining AI models with dedicated testing facilities and industrial partnerships aims to reduce that friction and create a scalable pipeline for future discoveries.

The funding round highlights a broader trend of AI companies moving beyond software applications into scientific research and industrial innovation. As demand for advanced semiconductors, energy systems, and sustainable manufacturing materials continues to rise, AI-assisted materials discovery is emerging as one of the most closely watched frontiers in deep-tech investing.