AINEWS
METI targets fiscal 2027 (April 2027) for a cross firm AI platform that pools R&D data to find new alloys, batteries, and chemicals, if rival manufacturers will share it.
Japan's Ministry of Economy, Trade and Industry (METI) wants to use AI to find new chemical compositions ten times faster. The agency is targeting fiscal 2027, which begins April 2027 in Japan's accounting calendar, for a cross-firm data platform that would let competing manufacturers pool proprietary R&D data and search for new materials using AI. The "new materials" in question are concrete industrial inputs: advanced alloys, battery chemistries, semiconductor materials, and industrial chemicals.
The tenfold speedup is a METI-stated target, not a measured outcome. It is also a deliberately round number. METI is choosing to aim for an order-of-magnitude acceleration in materials discovery, the kind of R&D that normally takes years of lab synthesis and testing per candidate. The bet is that AI trained on aggregated industrial data can short-circuit that loop. The harder question is whether rival firms will contribute the data at all.
Cross-firm data pooling has a long track record of stalling on intellectual property, antitrust review, and competitive concern. Materials R&D data is unusually sensitive: a company's recipe for a high-performance alloy or a battery cathode is often the closest thing it has to a trade secret. Sharing it with peers, even through a ministry-vetted intermediary, invites the worry that a partner firm will learn something it could later undercut you on. The platform's credibility will hinge on which firms actually contribute data and under what terms.
Mitsui Chemicals made a partial test of the model on July 28 when it joined the AI Materials Foundry as a founding member. The Foundry is a generative-AI network built around materials R&D, and Mitsui's involvement gives the consortium a credible industrial anchor. The Foundry and METI's planned platform are not the same thing. The Foundry is a consortium of firms working together on AI-driven materials discovery; METI's platform is a government-orchestrated data-sharing infrastructure. Conflating them risks reading a corporate partnership as a policy commitment.
METI is not approaching this from scratch. The ministry has been funding AI-ready manufacturing data and data-ecosystem work under its GENIAC project, including R&D themes announced in May and July 2026. That gives the platform a procurement clock. METI is not just publishing a vision, it is signing contracts and awarding grants against the fiscal-2027 launch date. The ministry is choosing to commit a deadline, and the deadline itself is the news the Nikkei scoop captured.
For Japanese industrial policy, the broader context is familiar. The country has spent two decades trying to coordinate manufacturer data in areas from automotive components to semiconductors, with mixed results. Pre-competitive consortia, where firms pool generic, non-sensitive data, have generally worked. Sharing data that touches a company's product roadmap has generally not. METI's design choices on what data sits in which tier will be the test.
A platform that launched with three firms contributing narrow, pre-competitive data, like generic materials science, public literature, and low-sensitivity process parameters, would be a useful but limited resource. A platform that attracted deep, proprietary data on the battery and semiconductor materials that anchor Japan's industrial policy could become a coordination mechanism the country has not had before. The next read is the participation roster when METI publishes it: who signs on, what they contribute, and what the published terms actually let other members do with the data.
The Nikkei scoop turned a long-running METI planning thread into a deadline. The fiscal-2027 date is news. The next read is the platform's participation roster and the legal terms METI publishes alongside it.