Cambridge-based AI startup CuspAI is reportedly in talks to raise at least $400 million in a funding round backed by Jeff Bezos’s family office, in a deal that would push the company’s valuation to $2.6bn.

According to a Financial Times report, investors in the round include Silicon Valley venture capital firm Kleiner Perkins and Bezos Expeditions, the family office of the Amazon founder. Term sheets have reportedly been signed, though the deal has not yet formally closed.

The financing would represent a meteoric rise in valuation for the two-year-old company, which was valued at $520 million in September 2025. CuspAI has previously raised more than $220 million from investors including Singaporean sovereign wealth fund Temasek and venture capital firm NEA.

CuspAI applies genAI to materials science, using AI and molecular simulation to accelerate the discovery of new materials for industrial use. The company’s platform uses “inverse design”, working backwards from a material’s desired properties before testing potential performance and manufacturability through digital simulation.

The approach is intended to shorten a process that can traditionally take years, with applications across semiconductors, aerospace, automotive, energy, climate technology and water purification.

Founded in 2024, CuspAI was co-founded by Dr Chad Edwards, a chemist and DeepTech entrepreneur who previously scaled quantum computing unicorn Quantinuum, and Professor Max Welling, a machine learning professor at the University of Amsterdam and former Distinguished Scientist at Microsoft Research and VP of Technology at Qualcomm.

The company counts a number of high-profile advisers, including AI pioneers Geoffrey Hinton and Yann LeCun, former BP chief Lord John Browne, and Martin van den Brink, the former president and CTO of ASML.

Announcing its Series A in September 2025, CuspAI CEO and Co-founder Chad Edwards said: “Next-generation AI compute, clean air and water and sustainable energy – these multi-billion-dollar global challenges share a common barrier: materials. CuspAI’s mission is to change that.”

CuspAI’s technology is being used by customers including chip equipment manufacturer ASML, Meta Platforms, and Hyundai to design and simulate new materials with specific properties. Its commercial partnerships span automotive, semiconductors, water purification, and climate technology.

The company has also worked with Finnish chemicals group Kemira on the development of materials designed to remove PFAS, often referred to as “forever chemicals”, from drinking water. Over a six-month project, the company’s systems reportedly narrowed 300tn potential material structures down to 20 candidate designs, which are now moving into further development.

The potential relevance of CuspAI’s work extends into digital infrastructure, where the rapid buildout of AI systems is increasing demand for advanced chips, cooling systems, thermal materials and power conversion hardware.

Recommended reading

• AI-powered Phishing Attacks Hitting Inboxes Every 19 Seconds
• ICO Sets Out Five Steps to Tackle AI Cyber Threats
• GCHQ Announces New National AI Cyber Shield

As AI infrastructure grows, operators are also facing greater scrutiny over water use, particularly in regions where data centres rely on evaporative cooling. More than 40% of planned and existing data centres in the United States are said to be located in areas classified as high or extremely high water scarcity, while projects worth an estimated $156bn were delayed or cancelled in 2025 following local resistance, with water concerns described as the largest category of community complaints.

The ability to accelerate the discovery of new filtration and water-treatment materials could therefore have commercial relevance for data centre operators, particularly where water access and permitting have become more significant constraints on development.

CuspAI was named as one of the frontier AI companies in the UK government’s AI for Science strategy and was among the first companies to gain significant access to the Isambard AI supercomputer