CuspAI, a two-year-old British startup, has secured $450 million in Series B financing to accelerate its artificial intelligence-driven approach to discovering new semiconductor materials.
The Cambridge, UK-based company launched the AI Materials Foundry on Monday, a coalition of more than 48 technology giants, industrial firms, and research facilities united around a shared goal.
Members of the coalition include Nvidia Corp. (NASDAQ: NVDA), Meta Platforms Inc. (NASDAQ: META), and Hyundai Motor Group, who will pool computing and scientific resources to develop new materials faster and at lower cost.
The funding round was led by venture firms Kleiner Perkins and NEA, with “significant participation” from Bezos Expeditions, according to CuspAI, valuing the startup at $2.6 billion.
That valuation represents a fivefold increase from just $520 million last September, underscoring the intense investor appetite for AI applications in materials science.
Chad Edwards, CuspAI’s co-founder and chief executive, said much of the new capital will support foundry labs in Cambridge, Singapore, and the San Francisco Bay area.
Edwards said chipmakers and semiconductor suppliers are “frantically searching for new materials,” noting his company has been “literally pulled by all four limbs” by demand from the chip manufacturing supply chain.
One of CuspAI’s stated goals is to reduce or eliminate reliance on rare metals with supply-chain risks, such as ruthenium and iridium, which are currently used in chipmaking processes.
Geetika Gupta, a senior director at Nvidia, said the chipmaker is seeing huge demand for new materials across its entire market, from energy storage systems to data center cooling.
Gupta said Nvidia will most likely pursue novel materials through “three-way collaborations” between CuspAI and a third partner, noting that “all of these companies, they treat a lot of this information as a secret sauce.”
CuspAI originally focused on finding new materials for carbon capture and water purification before pivoting over the past year in response to surging demand from the semiconductor industry.
An early project with Meta researchers involved building a library of 300 trillion possible structures within a class of chemicals called metal organic frameworks, which are particularly effective at capturing carbon dioxide.
That list was eventually narrowed to 10 candidates, but CuspAI was only able to synthesize fewer than 10 of those chemicals, and none outperformed what was already commercially available, Edwards acknowledged.
The startup is now applying the same library to identify materials capable of removing toxic “forever” chemicals from water, with Finnish chemicals company Kemira Oy set to synthesize and test 20 new structures this year.
CuspAI is trying to differentiate itself through scientific credibility, with co-founder Max Welling serving as a respected AI researcher and AI pioneers Yann LeCun and Geoffrey Hinton sitting on its advisory board.
The company also named Abhi Talwalkar, a semiconductor veteran and director of Advanced Micro Devices Inc. (NASDAQ: AMD), to its board, with AMD’s venture unit participating in the latest financing round.
John Giannandrea, a former Google and Apple executive, was hired to work part-time establishing CuspAI’s California outpost, adding further pedigree to its growing leadership team.
David Fairen-Jimenez, professor of molecular engineering at the University of Cambridge, acknowledged that machine learning has “certainly” sped up materials development but cautioned against overstating its potential.
“It’s super useful,” said Fairen-Jimenez. “What I don’t agree with is that there will be some magic coming from machine learning alone.”
Welling, for his part, remains focused on reducing the friction around real-world experimentation, arguing that insufficient data, testing facilities, and computing power have long held the field back. “People underestimate the friction of actually doing the experiment,” he said. “It feels maybe simple, but it’s not at all.”
