Cambridge researchers help build AI’s materials pipeline

demoduck.co.uk editorial team 2026-07-28 12:10 3 min read

Industrial machine testing a wooden beam for structural integrity in a laboratory.

More than 45 organisations are joining forces to tackle a bottleneck that could shape the next generation of semiconductors, clean-energy technologies and advanced manufacturing: finding materials that can be designed by AI and then made successfully in a laboratory.

Researchers from the University of Cambridge are contributing to the AI Materials Foundry, a global network launched by Cambridge-based start-up CuspAI. The initiative links high-quality data, advanced computing, experimental laboratories and scientific expertise across the full materials-discovery process.

Cambridge team connects AI predictions with laboratory tests

The Cambridge contribution comes from the Department of Materials Science & Metallurgy, where researchers are developing robotic platforms for high-throughput experimentation and advanced characterisation.

Led by Dr Shijing Sun, the team is working to connect computer-based materials design with automated synthesis, testing and rapid experimental validation. That connection is intended to show whether a material predicted by an AI model can actually be produced and assessed in practice.

Sun said the Foundry could help “close the loop between AI and experiments”, allowing academic and industry researchers to work together on materials with potential technological applications.

Cambridge researchers help build AI’s materials pipeline

Why computer-generated materials still need testing

Many generative AI systems used in materials discovery are currently assessed mainly through computer simulations. Those simulations can identify promising candidates, but a predicted material may never be synthesised or tested in a laboratory.

That gap limits the number of discoveries that can move beyond a digital prediction. The Foundry’s approach is designed to bring experimental results back into the AI process, helping future models focus on materials that are more likely to be successfully made and perform as expected.

The work is relevant to industries that depend on new or improved materials, including semiconductor production, clean energy and advanced manufacturing. The source does not claim that the network has already delivered a commercial material or solved those industries’ supply challenges; its stated role is to accelerate research and validation.

A 45-member network spanning research and industry

The AI Materials Foundry includes more than 45 members from academia, industry and research organisations. Participants named in the announcement include NVIDIA, Meta Platforms and the Henry Royce Institute, the UK’s national institute for advanced materials research and innovation.

The University of Cambridge is a founding partner of the Henry Royce Institute. Cambridge-based founding members of the Foundry also include the Cambridge Crystallographic Data Centre, reflecting the area’s combination of artificial intelligence, materials science, scientific data and experimental research expertise.

CuspAI CEO and co-founder Dr Chad Edwards said the network aims to combine agentic AI, domain expertise, data access and collaboration with customers to address the need for materials that do not yet exist.

Cambridge project will put generative AI candidates to the test

Sun is also developing a related project with Professor Aron Walsh of Imperial College London, who is CuspAI’s chief scientific officer. The project has received funding from the Alchemy Frontier Fund and will use generative AI to identify materials that can then be tested experimentally.

Its next stage is centred on making laboratory results part of the prediction cycle. Instead of treating simulation as the endpoint, the team plans to use experimental evidence to improve later predictions and concentrate research effort on candidates with a stronger chance of being realised.

Source: University of Cambridge Research News

Context & actions About this article

Source check Source trail

This report follows the supplied University of Cambridge Research News source and separates confirmed research plans from longer-term potential.

Checks

Confirmed the Foundry’s membership figure and named organisations
Checked the Cambridge department and research roles described
Separated laboratory validation plans from claims of completed commercial results

University of Cambridge Research News Check facts