From $520M to $2.6B in 9 months: Bezos Expeditions co-leads CuspAI’s $450M Series B to scale AI materials discovery — TFN

• CuspAI raised $450 million in a Series B round at a $2.6 billion valuation, led by Kleiner Perkins and NEA, with participation from Bezos Expeditions.
• The Cambridge-based startup’s valuation has grown almost five times in nine months, rising from $520 million to $2.6 billion, and it has now raised more than $650 million in total.
• ASML, Meta, Samsung, and Hyundai already use CuspAI’s materials search engine, and the company has simultaneously launched the AI Materials Foundry, a network of over 45 partners, including NVIDIA and Meta.

CuspAI, a Cambridge, UK-based startup, has closed a $450 million Series B at a $2.6 billion valuation, led by Kleiner Perkins and NEA, with participation from Jeff Bezos’s family office, Bezos Expeditions.

New investors include Glade Brook Capital Partners, Lux Capital, AMD Ventures, Tru Arrow Partners, StepStone, Britain’s Sovereign AI Venture Fund, the Netherlands’ Invest-NL, and John Doerr, alongside returning backers Temasek, Basis Set Ventures, Giant Ventures, Touring Capital, Prosus, Phoenix Court, and Northzone.

“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. What sets Max, Chad, and the team apart is that they design for materials that can actually be built, not just ones a model can dream up,” says Josh Coyne, partner at Kleiner Perkins.

CuspAI was founded in 2024 by Dr Chad Edwards, a chemist who helped grow Quantinuum, and Professor Max Welling, a machine learning expert from the University of Amsterdam and former Distinguished Scientist at Microsoft Research who co-invented the variational autoencoder architecture underpinning generative molecular design.

The company has about 65-68 employees and now operates across Cambridge, London, Amsterdam, Berlin, Tokyo, Singapore, and the United States, including a London King’s Cross office.

The global market for AI-driven materials discovery was worth about $0.74 billion in 2025 and could reach $2.77 billion by 2030, growing at a 30% annual rate. This rapid growth is driving investments from CuspAI, ASML, sovereign wealth funds, and Bezos.

CuspAI describes its platform, called MIRA, as a search engine for the material world. A partner enters the physical properties they need, such as strength, conductivity, or thermal tolerance, and the system suggests matching chemical compositions, designs synthesis routes matched to available lab infrastructure, and routes the work to the right facility.

The company claims this process is up to ten times faster than traditional trial-and-error research. For instance, a semiconductor manufacturer looking for a heat-resistant conductor can generate and test candidate compounds on the platform before starting lab work. With Finnish chemicals company Kemira, CuspAI says it screened a search space of 300 trillion potential molecular structures and delivered twenty validated candidates in six months, a process that previously took years.

“As AI transforms the physical world, new materials will open up new frontiers across semiconductors, energy and advanced manufacturing,” says Ian Buck, vice president of hyperscale and HPC at NVIDIA, which is supplying compute infrastructure to the network.

Rob Fergus, vice president of AI research and head of FAIR at Meta, whose team built the Universal Model for Atoms used in the platform’s simulation layer, adds: “Our open source frontier models for materials science research will enable more teams to tackle previously intractable challenges, more precisely and more quickly than before.”

CuspAI stands out in a crowded market by working with real customers

CuspAI is one of several companies using AI to speed up materials research and development.

Its main competitors are Orbital Materials, a spinout from DeepMind alumni, and XtalPi, which is valued at $2.5 billion. Other established players include Dassault Systèmes and Schrödinger.

The competition is even tougher with Flagship Pioneering’s Lila Sciences, which has raised over $550 million, and Periodic Labs, started by former OpenAI and Google DeepMind staff with a $1 billion valuation. European competitors include Altrove in Paris, Dunia Innovations in Germany, and Molecular Quantum Solutions in Denmark.

CuspAI sets itself apart by focusing on customer adoption rather than technology claims. Its clients include ASML, Meta, Samsung, and Hyundai. The company has also worked with Hyundai on sustainable energy projects and with Kemira to remove PFAS compounds from water.

From a $30M seed to a $2.6B valuation

The Series B comes after a $30 million seed round in June 2024 and a $100 million Series A in September 2025, co-led by New Enterprise Associates and Temasek at a $520 million valuation.

New commercial contracts had already pushed the valuation to about $800 million before this round. CuspAI has now raised more than $650 million in total. The latest funding will help scale computational infrastructure, grow materials research teams, and bring the molecular design platform to market.

“Since NEA co-led the Series A, we’ve seen Chad and Max’s relentless focus and ability to build a world-class team. We’re proud to double down on our investment as CuspAI transforms how the world’s most vital industries design and deploy next-generation materials,” says Philip Chopin, managing director and head of Europe at NEA.

Bezos Expeditions has invested in companies such as Anthropic, Perplexity AI, Physical Intelligence, and Figure AI, typically taking smaller minority stakes rather than leading rounds. This Series B closed soon after Bezos announced Prometheus, his $41 billion physical AI lab. Only Bezos can say whether the timing was planned, but the two projects seem closely linked: one lab designs physical systems, while the other supplies the materials to build them.

“Materials discovery has quietly been a bottleneck behind progress in energy, semiconductors, and climate. What’s remarkable about Cusp is that a single AI platform is already solving hard problems across industries — from stripping chemicals forever from drinking water to designing next-generation chip materials,” says John Doerr, of Doerr Capital, a new investor in this round.

Edwards frames the mission bluntly: “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. That’s what we’re on a mission to solve”

The next 18 months will show whether CuspAI’s models can really cut research cycles from decades to months across more than 45 Foundry partners at once, or if this is another case where AI hype outpaces real lab results.