Singapore puts $150m into AI that tests 470 alloy samples a day
Singapore is betting about $150 million of government money that artificial intelligence can shorten the years it takes to turn a promising material into a manufacturable product. Deputy Prime Minister Gan Kim Yong launched MARVEL-AI at the Singapore Week of Innovation and Technology, tasking the platform with judging whether new alloys, polymers and coatings can be made at scale and sold commercially.
The gap it targets is the long stretch between discovery and deployment. According to Enterprise Singapore, validating a material's properties, testing it under real conditions and proving it can be manufactured takes multiple development cycles and repeated handovers between research teams and production engineers. MARVEL-AI is designed to link those stages: AI models will propose candidate materials and steer experiments, automated laboratories will make and test samples, and the results will decide what researchers examine next.
The throughput case rests on existing laboratory work rather than anything achieved since the launch. Enterprise Singapore said one alloy process can already make and test up to 470 different samples a day, more than 400 times the rate of conventional methods. Professor Loh Xian Jun, executive director of MARVEL-AI, said the platform "builds on Singapore's established strengths in materials science and connects them with AI, automated experimentation, manufacturing and industrial translation", bridging the gap between discovery and deployment.
Industry partners have signed on around specific manufacturing needs. ST Engineering is working on high-performance alloys for aerospace, defence and infrastructure; Heraeus Electronics on powders and alloys for manufacturing and electronics; and Dura-Metal on protective coatings and cladding alloys for harsh environments. Foxconn Industrial Internet will pursue sustainable materials and printable alloys intended to reduce reliance on critical raw materials, while Suntar Research focuses on advanced membrane materials.
Technology partners will supply the AI layer. CuspAI will connect its generative system to the automated laboratories to help design materials, and Google Cloud will build a secure research environment giving materials scientists early access to its advanced capabilities. The platform is anchored by A*STAR's Institute of Materials Research and Engineering, with A*STAR, Enterprise Singapore and the Economic Development Board connecting companies and investors to its equipment, testing capabilities and researchers.
