Singapore × A*STAR's MARVEL-AI × five industry partners: all five roles designed into one platform

One of the Buyer Chains series. Each piece follows one chain, the AI company, its customer and sometimes the customer's customer, through five roles in the public record: who states the need, who sets the research direction, who runs the experiments, who validates the result, who decides to keep paying. It records only what the public record states and says who said each thing; a missing step is marked not yet public rather than filled in, and a company's own figure is written as the company's. The timeline at the end stays open: when the chain's next step reaches the public record, it goes on as a new line.

On 27 October 2026 Singapore's Deputy Prime Minister Gan Kim Yong announced MARVEL-AI at SWITCH 2026: a materials platform led by A*STAR, starting with S$150 million of government money, in which AI models pick candidates and design experiments and automated labs make and test them.[1] A*STAR's release the same day gives the structure: anchored at its Institute of Materials Research and Engineering (IMRE), supported by the trade ministry, EnterpriseSG, EDB and the National Research Foundation under RIE2030, covering four material classes (metal alloys, polymers and organics, ceramics, composites) for semiconductors, advanced manufacturing, energy and chemicals; it names five industry partners, each with a material need, and five technology partners, among them CuspAI, XtalPi, ChemLex, Google Cloud and Shimadzu.[2] The release says the point plainly: "discovery alone is not enough. The next challenge is proving they can work in the real world" (Loh Xian Jun, MARVEL-AI's executive director).[2] What is distinctive about this chain is that it was designed: each of the five roles has a place in the announcement, and no material yet sits in any of them.

What the buyer looks at, concretely

• All five partners' needs are directions, not numbers. The acceptance standards that match them come from each industry: aerospace alloys go through AMS specifications and MMPDS design allowables; electronics powders and alloys against the thermal and mechanical targets of the packaging roadmap (IEEE HIR, 2025 edition); membranes on flux, selectivity and lifetime.[3] Which of these becomes MARVEL-AI's first acceptance test waits on the first project.
• The platform's own promised steps can be counted: the first material a partner tests to its specification; the first process taken to "manufactured at scale"; the first company that locates or expands because of it (the release lists "anchor large multinationals with materials development mandate" as a goal).[2]
• For a technology partner such as CuspAI, what the platform gives is lab access, not procurement: its role here is to connect its models to someone else's lab, and the buyer sits at the platform's other end.

Not yet public · The S$150 million's allocation and term; each partner's money or commitment; the first project, material and acceptance test; the baseline and tester behind the capsule example.

Decision point · The next decision is A*STAR's and the five industry partners': which material class the first project takes, and whose specification it is accepted against. The platform has a place for all five roles; this decision is not public.

The chain's timeline

• 2026-10-27The Deputy Prime Minister announces MARVEL-AI with S$150 million (speech)[1]; A*STAR's release names five industry and five technology partners (release)[2]
• NextThe first named project: which partner, which material class, which metric. Awaiting public data.
• LaterThe first material a partner adopts; where the S$150 million goes.

All cases →