The development of new materials is critical in the efforts to deploy clean energy, accelerate advanced manufacturing and in the development of semiconductors. In a bid to step up the rate of discovery of new materials, CuspAI has launched the AI Materials Foundry, seeking to remove a bottleneck it said is constraining progress across semiconductors, clean energy and advanced manufacturing.
The AI Materials Foundry comprises 45 founding members from around the world. These include Nvidia which is providing the compute infrastructure, and Meta’s Fundamental AI research team developing the Universal Model for Atoms platform, described as a frontier chemistry model for materials science. Other members include Johnson Matthey, Merck, Mitsui Chemicals, Topsoe, along with a laboratory network which includes the Henry Royce Institute, Cambridge University and the University of Amsterdam.
Centred around CuspAI’s proprietary AI platform called MIRA, the partners will run full discovery cycles – from generative materials design through to simulation, developing routes to synthesise the new materials, and coordinating experimental validation.
The AI Materials Foundry will bring together for key elements: training data at scale to make AI predictions reliable; computing power that will allow screening of billions of candidates at molecular resolution; infrastructure to move from digital design to physical reality; and expertise to interpret what AI finds and what to do with it. “Rather than relying on the physical constraints of a single isolated laboratory, CuspAI’s ecosystem approach means breakthroughs achieved within the network have the potential to accelerate discovery timelines across the entire global value chain,” CuspAI said.
CuspAI adds that it has secured exclusive access AI training rights to the datasets that make up the foundational experimental records of materials science, including the Cambridge Structural Database and the Inorganic Crystal Structure Database. It also has licenced access to materials science content from Wiley and other leading publishers of scientific journals.
Professor Paul Monks, Chair of the Henry Royce Institute one of the founding partners based in the UK said: “Our ambition is to embed AI across the entire materials value chain from discovery and characterisation, through scale-up, manufacturing, deployment and recycling. By connecting world-leading research capability with emerging AI technologies, we can accelerate innovation and help deliver advanced materials needed for a more sustainable and resilient future.”
Omote Toshihiko, CTO and Managing Executive Officer at Mitsui Chemicals added: “We view [our participation] as an opportunity to structurally strengthen our competitive advantage. We expect AI-driven materials discovery to significantly improve the speed of R&D and the success rate of new business creation, generating tangible business value.”
Further reading
- Hydrogen-powered flight: Filling the gaps in materials knowledge
- Lithium-ion batteries: New recycling method restores spent cathode materials
Visit Blog