Singapore funds $10 million artificial intelligence lab to solve materials manufacturing bottlenecks in chips and clean hydrogen

Singapore is putting $10 million behind a joint laboratory designed to turn AI and automation into manufacturing recipes for critical materials-the kind of recipes chipmakers and clean hydrogen producers need before they can scale. The Materials Data Foundry, a collaboration between the National University of Singapore and the University of Toronto's Acceleration Consortium, sits inside Singapore's national AI-for-Science programme, which the National Research Foundation is funding with $120 million in total.

The AI-for-Science programme, or AI4S, launched in October 2024, but the named projects, including the foundry, show where the city-state is directing money: at the gap between a promising lab result and a factory-ready process. The announcement came alongside the AI4X Accelerate Conference at Raffles City Convention Centre, a five-day event that drew more than 800 researchers, engineers and industry participants, according to organisers. Self-driving laboratories and machine learning for scientific discovery anchored the agenda.

Semiconductors hit a materials wall

As transistor scaling enters the angstrom era, the list of usable dielectrics, conductors and etch chemistries gets narrower. Traditional trial-and-error discovery cannot keep pace with the roadmaps chipmakers are chasing. Singapore wants to build capability close to home. Applied Materials already opened a $644 million manufacturing plant in Singapore and has deepened its research collaboration with NUS. Beginning in August 2026, NUS will introduce an Applied AI for Materials and Process Engineering specialisation inside its Master of Science in Semiconductor Technology and Operations programme. That course signals the skills gap: the next useful hire may need to understand both machine learning and materials chemistry.

A lab that helps shorten the path from candidate material to manufacturable process is not glamorous. But it is the kind of infrastructure that keeps a country embedded in the semiconductor supply chain.

Clean hydrogen's manufacturing gap

Clean hydrogen research faces a parallel constraint. NUS runs a Centre for Hydrogen Innovations that uses computational catalysis and machine learning methods-including density functional theory-for water electrolysis. The problem is rarely the discovery itself; it is making a promising catalyst repeatedly, at scale, without the economics collapsing. The Materials Data Foundry is aimed at that middle stage: testing thousands of AI-guided experimental paths and converting the best ones into production instructions.

The University of Toronto's Acceleration Consortium contributes its self-driving laboratory model, where AI designs experiments, robotic systems run them, and the results feed the next round of decisions. Similar automated approaches have worked in drug discovery. Recursion Pharmaceuticals screened huge compound libraries far faster than a conventional human-led workflow. Inorganic materials chemistry introduces different challenges, but the logic is the same: treat experimentation as a learning loop, not a slow queue.

Why this matters for science and research professionals

The foundry's ambition is not to produce a commercial catalyst next quarter. It aims to make automated materials discovery reliable enough that companies trust the recipes and build around them. For researchers in materials science, chemistry and adjacent fields, the project highlights where public funding is trying to shrink the gap between discovery and manufacturing. The ability to work with AI-led experimental design and self-driving labs is becoming a practical skill, not a niche interest. Singapore's bet is that the people who can bridge machine learning and materials chemistry will anchor the next phase of industrial research in the country.