CuspAI, an AI materials-discovery startup headquartered in Cambridge, England, has announced the launch of its AI Materials Foundry, a global network that combines the expertise of more than 45 partner organizations to discover useful novel materials.
The founding group includes NVIDIA, which will provide the compute infrastructure, and Meta’s Fundamental AI Research Team, who develop the Universal Model for Atoms (UMA), a frontier atomistic chemistry model for materials science.
The AI Materials Foundry will focus on discovering novel semiconductors and other materials useful in clean energy and advanced manufacturing. The company says the AI Materials Foundry delivers four key attributes necessary for successful software-led materials discovery:
“The world needs materials that don’t yet exist,” said Dr. Chad Edwards, CuspAI founder and CEO, in a statement, positing that the next 50 years of industrial progress could be constrained by a lack of novel materials without fast progress in making discoveries. “That’s what we’re on a mission to solve – combining frontier agentic AI with deep domain expertise, exclusive data access and close customer partnerships,” he added. “We’re delighted to be joined by more than 45 leaders in their fields to advance materials discovery.”
Among the partners are companies with ties to the solar industry, such as perovskite-focused Caelux and Oxford PV, encapsulant maker Mitsui Chemicals and semiconductor process technology provider Applied Materials. Also joining are 3M and Fujifilm, companies with broad product portfolios that include solar materials.
“We believe AI-driven materials discovery can significantly accelerate the path from scientific insight to commercial impact,” said Oxford PV chief technology officer Ed Crossland. “Through the AI Materials Foundry, we look forward to contributing our expertise in advanced photovoltaic materials and collaborating with a world-class network of partners to accelerate innovation in clean energy.”
In a statement provided to pv magazine USA, Caelux CEO Scott Graybeal said the following:
“Caelux is honored to be a founding member of CuspAI’s ‘AI Materials Foundry’ with other leading technology companies. As we continue to develop the lowest cost form of power generation with our perovskite solar technology access to the world’s leading data set will accelerate our efforts. Our own R&D has shown how different material combinations can improve perovskite’s performance. As we enter product commercialization this will further increase our technology advancement, delivering on our promise to provide more power, more energy and lower LCOEs for our customers.”
How the CuspAI platform works
CuspAI will use its MIRA AI platform to orchestrate the process of discovering new materials using what the company calls “the largest curated experimental materials datasets in the world.”
The discovery platform can be deployed as a private instance within a company’s existing R&D process, with MIRA acting as an autonomous AI agent to generate candidates for simulation and testing from a list of attributes partners share about the materials they’re seeking.
The materials and their physical properties can be simulated by an open source molecular simulation toolkit called kUPS, which was built by CuspAI in collaboration with the NVIDIA ALCHEMI (AI Lab for Chemistry and Materials Innovation) team. The tool uses Meta’s UMA to simulate atomic interactions across the periodic table.
As an example of one successful deployment of its technology, CuspAI points to a project it conducted for Finnish chemicals company Kemira, in which its system screened as many as 300 trillion potential molecular structures to find twenty candidates for further testing and validation. The company said this allowed Kemira to accomplish in six months what would have taken it years under its previous process.
Clean energy applications for AI
The practice of discovering novel materials using AI has grown rapidly in recent years. One notable initiative is Google’s Graph Networks for Materials Exploration (GNoME), which has discovered more than 380,000 highly stable materials to date (and released a dataset of 520,000 to the public). Its discoveries include thousands of new layered compounds such as lithium-ion conductors that could improve battery performance.
Microsoft’s MatterGen system — a generative diffusion model for inorganic materials design — does something similar to what the CuspAI system promises, generating novel atomic structures based on the specific properties requested by users.
In the solar industry, scientists at UC San Diego and the University of Southern Denmark have used machine learning to develop and test novel perovskite materials, furthering efforts to create perovskite solar cells that remain stable over long periods of exposure to real-world conditions.
Outside of materials discovery, machine learning is also used to study the physical properties of deployed solar installations, allowing for more efficient and effective operations and maintenance practices.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: [email protected].
Comments
Please login to comment
Tuesday, August 11, 2026
3:00 pm – 4:00 pm CEST, Berlin, Paris, Madrid
Thursday, July 30, 2026
4:00 pm – 5:00 pm CEST, Berlin, Paris, Madrid
Thursday, July 16, 2026
4:00 pm – 5:00 pm CEST, Berlin, Paris, Madrid
The June issue of pv magazine Global is out now!
Available in print and digital – get your copy today!
Thursday, October 7, 2026
11:00 am – 12:30 pm CEST, Berlin, Paris, Madrid
A two-day conference in Austin, Texas, bringing together leaders in US solar manufacturing, equipment specification, and factory execution.
Entries open in seven categories: Modules, Inverters, BoS, BESS, Manufacturing, Sustainability, Projects.
April 01 – August 31, 2026
pv magazine USA hosts its third multi-day virtual event on advancing U.S. solar and energy storage markets, covering financing, supply chains, and distributed energy’s role in grid resilience.
You have no items in your basket.
