TECHNOLOGY EXPLAINER
National MI Strategies and Data Infrastructure
- How to read "X billion invested" and "national strategy" (the US, Europe, Japan and China, and the terms of access to data infrastructure)
When materials informatics (MI) is in the news, we often see phrases such as "a government invests tens of billions of yen" or "a national strategy to speed up materials development with AI". But what that figure means changes greatly depending on whether it is a plan, a call for proposals, a contract or money actually spent. And even the word "strategy" covers documents of quite different natures: an executive order, a policy communication, a government decision, a funding guideline and an expert recommendation. This article assumes a materials engineer new to MI. It explains (1) the stages behind the words used for money, (2) what the US, Europe, Japan and China state in their official documents, and (3) who can use each country's data infrastructure, at what price and how. We do not rank countries. Amounts are given in the original currency, as worded in the documents, with a checking date (23 October 2026).
- The short version / before you start / where this fits in the series
- Reading the words "budget" and "strategy"
- The United States: MGI and the Genesis Mission
- Europe: from a research framework to a common data infrastructure
- Japan: strategy, the Materials DX Platform and NEDO
- China: Materials Genome Engineering and "AI Plus"
- A timeline, and why not to compare
- Terms of access to data infrastructure: a materials engineer's view
- Limits, what we could not confirm / summary
- How to read the yardsticks and standards
- Glossary
- References (primary sources)
- Claim-to-source audit
Sourced = stated in a published source (link given)
Our calculation = a value this article computed from stated assumptions (in this article, mostly our organization of sources)
Not yet confirmed = plans, targets, proposals, schedules and other items not yet settled
Our own readings are marked as commentary. Inference is shown in bold.
Amounts differ in currency, fiscal year, period and stage. We do not convert exchange rates; we keep the original currency and the words of each document (proposal, estimate, indicative amount, contract, call and so on). Dates and status are as of the checking date (23 October 2026). This article does not rank countries, does not give advice on investment or policy toward any company or country, and does not give legal advice. Items that have not yet been proposed or enacted (the European Advanced Materials Act, the next European research framework, the Japanese expert recommendation and so on) are not treated as settled systems. We could not confirm the outcomes of the US Genesis Mission, independent verification of the targets of the Japanese NEDO project, the results of the Chinese special project or the amounts enacted for it, or amounts for the US MGI, so we do not state them.
1. The short version / before you start / where this fits in the series
- A "national MI strategy" is not one kind of thing: the US has a multi-agency strategic plan (MGI) and an executive order (the Genesis Mission); Europe has a policy communication, a research framework and joint projects; Japan has a Cabinet Office strategy, MEXT programmes and NEDO and JST projects; China has funding guidelines of the Ministry of Science and Technology and an opinion of the State Council. The nature of the document differs by country [References 2, 3, 6, 10, 16, 17].
- Amounts come in stages: for example, the US$293 million of the US Department of Energy (DOE) is the amount of a "call", the EUR 175 billion for the next European research framework is a "proposal", about 13.8 billion yen for the NEDO project is a "plan", and NEDO's actual contribution totals 13.751 billion yen Sourced [References 4, 5, 14]. The same "X billion" means different things.
- For a materials engineer, the terms of access to data infrastructure are a practical issue: published data in NOMAD (Europe) are under CC-BY-4.0, data sharing from Japan's ARIM is a paid annual subscription with eligibility limited to residents and corporations in Japan, and NIMS's MDR specifies a licence for each dataset Sourced [References 9, 12, 18]. "A public data infrastructure exists" and "your company can use it" are different things.
"Plan", "proposal", "call", "contract" and "actual" are different numbers. When you read an article on a national MI strategy, check which stage a figure belongs to and to whom the data infrastructure is open. The size of a national budget is not a yardstick of how far MI has progressed.
MI (materials informatics) is a way of advancing the search for and design of materials with materials data and machine learning. Searching for materials is like "looking for treasure on a vast map", and MI is a tool for narrowing where to look. Calculation "gives candidates cheaply and fast", AI "narrows where to look", and experiment "checks whether it is true".
Where national strategies touch materials, it is mainly through "data infrastructure". Data infrastructure can be compared to a "common atlas" for materials searching. Limits of the analogy: unlike an atlas, the data in these infrastructures differ in whether they are calculated or measured and in who measured them when and how, and who may use them and at what cost also differs by infrastructure.
This article is the policy article of the Materials Informatics series. Databases and the FAIR principles are covered in our explainer on materials databases and the FAIR principles, autonomous laboratories in our explainer on self-driving laboratories, and how to read evaluations in our explainer on how to read evaluation and reproducibility in MI. National strategies in other fields (space, quantum) are organized in the same way in our explainers on national space strategies and regulation and on national quantum strategies.
2. Reading the words "budget" and "strategy"
(1) Money comes in five stages
Unless you check which stage a figure in the news, "X billion invested", belongs to, you cannot tell what it means. In this article we divide money into five stages (our own organization). (1) A proposal is an amount that a government or commission has put forward as a draft for others to decide. (2) A plan or estimate is an amount written in a project plan. (3) A framework set is an amount decided as a budget framework (including indicative amounts). (4) A call or contract is the total of a call for a specific programme or an amount fixed in a contract. (5) The actual is the amount really paid.
A concrete example is in a DOE announcement. DOE announced a "US$293 million call" for teams working on the challenges of the Genesis Mission. Phase I awards range from US$500,000 to US$750,000 (nine months) and Phase II awards from US$6 million to US$15 million (three years) Sourced [Reference 4]. This is the frame of a call, not the result of selection. In addition, the text of the executive order says that work on the challenges is "subject to available appropriations" [Reference 3]. "Announced" and "available to spend" are different stages.
(2) The kind of document decides what can be decided
| Type of document | Examples in this article | What it decides | Cautions when reading |
|---|---|---|---|
| Executive order | US: Executive Order 14363 "Launching the Genesis Mission" (24 November 2025) [3] | Instructions, deadlines and responsibilities for executive agencies | No amount is written. The text says "consistent with applicable law and subject to available appropriations" Sourced |
| Strategic plan (multi-agency) | US: MGI Strategic Plan (November 2021) [2] | Goals, priorities and the direction of agency actions | No amounts were found in what we checked Sourced |
| Communication | Europe: COM(2024) 98 "Advanced Materials for Industrial Leadership" [6] | The policy of the European Commission and the direction of its actions | A policy document. Separate from a legislative proposal (the Advanced Materials Act) Sourced |
| Grant agreement | Europe: MaterialsCommons (grant agreement 101293545) [8] | The period and the amount of a grant | A contracted amount, not actual spending |
| Strategy (decision) | Japan: Materials Innovation Strategy (decided 27 April 2021, revised 4 June 2025) [10, 11, 27] | Government policy and actions to take | No amounts in the outlines we checked Sourced |
| Expert recommendation | Japan: Measures for implementing the Materials Innovation Strategy (12 June 2026) [27] | A proposal by an expert panel | A recommendation, not a government decision Sourced |
| Project record | Japan: NEDO project "Ultra-Advanced Materials Ultra-Fast Development Platform Technology" (title in Japanese; our translation) [14] | Project goals, the course of the budget and a self-description of results | Statements of achievement are the project side's own summary (independent verification not confirmed) |
| Application guideline / opinion | China: 2017 guideline of the Ministry of Science and Technology; State Council opinion (Guofa [2025] No. 11) [16, 17] | Direction of support, estimates and conditions for applying | The "estimate" in the guideline is a planning figure Sourced |
All Sourced (official documents, checked 23 October 2026). The arrangement of the table is Our calculation (organization). The legal effect of each document depends on its text and the responsible body (this is not legal advice).
In fact: the size of a budget is not a yardstick of how far MI has progressed. Results also depend on the quality and quantity of data, people, instruments and the strength of companies, not on money alone. And, as the figures of each country differ in scope, period, stage and currency, they cannot simply be lined up. This article does not rank countries (commentary).
The US federal fiscal year starts in October of the previous year and ends in September (for example, FY2026 runs from October 2025 to September 2026). The Japanese national "fiscal 2026" runs from April 2026 to March 2027. A research framework such as Horizon Europe is a multi-year framework (for example 2021 to 2027). Because the divisions differ by country, lining them up as "the budget of the same year" creates gaps (commentary).
3. The United States: MGI and the Genesis Mission
US efforts related to MI can be organized into two strands: the Materials Genome Initiative (MGI), which began in 2011, and the Genesis Mission, which began with an executive order in November 2025. They are separate frameworks.
(1) MGI: speeding up materials development with data and computation
The MGI website describes MGI as "a federal multi-agency initiative for discovering, manufacturing and deploying advanced materials twice as fast and at a fraction of the cost compared to traditional methods" Sourced [Reference 1]. The 2021 strategic plan states that MGI was launched in 2011 to accelerate the discovery, design, development and deployment of new materials, at a fraction of the cost, by using data and computational tools together with experiment, and that the 2021 plan updates the first plan of December 2014 [Reference 2].
The 2021 plan lists three goals for the next five years: (1) unify the Materials Innovation Infrastructure (MII) (connect computational, experimental and data tools), (2) harness the power of materials data (application to AI, and making data FAIR so that it is "AI-ready"), and (3) educate, train and connect the workforce [References 1, 2]. Under (2) are fostering a National Materials Data Network and data management plans for federally funded research (based on the FAIR idea) [Reference 2]. The plan cites the "Materials Innovation Platforms (MIP)" that the National Science Foundation (NSF) started in 2016 as an example of a national user facility built on the MGI approach [Reference 2]. In 2024 the "MGI Challenges" were announced [Reference 1].
What we could not confirm: in what we checked, the 2021 strategic plan gives no amounts per programme [Reference 2]. MGI is a framework spanning several agencies, and we could not confirm whether it has a single dedicated budget.
(2) The Genesis Mission: a national effort to speed up science with AI
Executive Order 14363, "Launching the Genesis Mission", was signed on 24 November 2025. Its text establishes the Genesis Mission, a national effort to accelerate scientific discovery with AI, with the Secretary of Energy responsible for implementation and the Assistant to the President for Science and Technology providing overall leadership Sourced [Reference 3]. The main contents are as follows.
- Platform: DOE establishes and operates the "American Science and Security Platform". It is to provide computing resources (national-laboratory supercomputers and secure cloud environments), AI frameworks and agents, predictive models, domain-specific foundation models, secure access to datasets, and tools for autonomous and AI-augmented experimentation and manufacturing.
- Deadlines: within 60 days DOE submits a list of at least 20 "national science and technology challenges"; within 90 days it identifies computing, storage and networking resources; within 120 days initial data and model assets (including digitization, standardization, metadata and provenance tracking); within 240 days it reviews robotic-laboratory capabilities; and within 270 days it seeks to demonstrate an initial operating capability of the Platform for at least one challenge ("consistent with applicable law and subject to available appropriations").
- Domains of challenges: advanced manufacturing, biotechnology, critical materials, nuclear fission and fusion energy, quantum information science, and semiconductors and microelectronics.
DOE's Genesis Mission website states the goal as "doubling the productivity and impact of American research and innovation within a decade". For materials it gives examples such as "securing critical materials" (co-designing substitutes and recovering rare elements from waste) and "discovering mission-ready materials" (alloys, polymers and composites, combining simulation, literature mining and autonomous laboratories) Sourced [Reference 20].
A DOE announcement of March 2026 presented a US$293 million call (an RFA, "The Genesis Mission: Transforming Science and Energy with AI"; NOFO DE-FOA-0003612). It covers more than 20 national challenges in advanced manufacturing, biotechnology, critical materials, nuclear energy, quantum information science and other areas. Phase I (nine months) awards are US$500,000 to US$750,000 each and Phase II (three years) awards US$6 million to US$15 million. Phase I applications and Phase II letters of intent were due on 28 April 2026, and Phase II applications on 19 May Sourced [Reference 4].
What we could not confirm: the results of selection and the amounts awarded under this call, the amount of the budget of the Genesis Mission as a whole, and how far the deadlines in the executive order (60 days, 90 days and so on) have actually been met. The US$293 million is the amount of the call, not the budget of the whole Genesis Mission Not yet confirmed.
An executive order is a document that tells executive agencies what to do, and it contains no amount. Money can actually be spent only after Congress enacts a budget. The text of the order itself says "subject to available appropriations" [Reference 3]. "The Genesis Mission was started by an executive order" and "X billion dollars can be spent" are different statements (commentary).
The MGI plan says that making data FAIR is a foothold for making data "AI-ready", and the Genesis Mission order lists digitization, standardization, metadata and provenance tracking for data assets [References 2, 3]. From a materials engineer's side, this points in the same direction as leaving measurement conditions, sample history and failed conditions in your lab notebook in machine-readable form (commentary). Whether or not you use or join a US infrastructure, how you organize your own data connects to any future infrastructure.
4. Europe: from a research framework to a common data infrastructure
The European Union (EU) is a group of 27 Member States, and efforts on materials overlap: EU-wide research frameworks, policies of the European Commission, joint projects and the initiatives of each Member State.
(1) The research framework: Horizon Europe
Horizon Europe is the EU's main research and innovation funding programme. According to the Commission's page, following the decision of the mid-term review of the Multiannual Financial Framework, the indicative funding amount for 2021 to 2027 is EUR 93.5 billion. The same page shows that on 16 July 2025 the Commission proposed a budget of EUR 175 billion for the next framework, 2028 to 2034, and that negotiations among the European Parliament, the Council and the Commission will follow Sourced Not yet confirmed [Reference 5]. The EUR 93.5 billion is an "indicative framework" and the EUR 175 billion is a "proposal", which has not yet been enacted.
(2) Policy: the Communication COM(2024) 98
On 27 February 2024 the European Commission issued the Communication COM(2024) 98, "Advanced Materials for Industrial Leadership" Sourced [Reference 6]. It states that developing advanced materials with conventional methods can take 10 to 30 years, that digitalisation of R&D could accelerate the discovery of innovative materials, and, as an example, that AI recently helped to predict nearly 400,000 stable crystal structures. As an action, it sets as a goal that the Commission and Member States build, by mid-2025, a long-term sustainable European digital infrastructure for advanced materials R&I, the "materials commons", and says they will explore launching a European Digital Infrastructure Consortium. This infrastructure is to link with the European Open Science Cloud (EOSC), European data spaces such as the Manufacturing Data Space, national initiatives (MaterialDigital, Diadem) and the EU project BIG-MAP, and to be based on the FAIR principles [Reference 6].
How to read it: the "nearly 400,000 stable crystal structures predicted" in the Communication is a count of computed predictions (chapter 7 of our explainer on how to read evaluation and reproducibility in MI covered how to read this kind of "discovery"). That is, a policy document cites a count of predictions as an example of the possibilities of AI (commentary).
(3) Law: the Advanced Materials Act (planned)
On 21 October 2025 the European Commission launched a call for evidence and a public consultation on a future Advanced Materials Act. Contributions were accepted until 13 January 2026. The Act is described as supporting the research and innovation process from design and development to manufacturing and deployment, tackling the long time from design to deployment and the expansion of production capacity of advanced materials in the EU, and streamlining regulatory processes, and it is expected to be proposed in 2026 Sourced Not yet confirmed [Reference 7]. We could not confirm a record that a proposal had been issued as of the checking date. This article does not describe the content of the Act as a settled system.
(4) A joint project: MaterialsCommons (started June 2026)
A project close to the "materials commons" of the Communication, MaterialsCommons, has started with a Horizon Europe grant (grant agreement 101293545). According to the CORDIS record, the EU signature date is 12 May 2026, the project period is 1 June 2026 to 31 May 2030, and the total cost is EUR 27,998,554.91, with the EU contribution of the same amount Sourced [Reference 8]. According to an announcement by a participating institution (2 July 2026), it is a consortium of 26 academic partners, 21 associated partners and 12 associated collaborators, coordinated by the Fraunhofer Institute for Mechanics of Materials (Germany), and its kick-off was held on 17 June 2026. Its aim is to connect existing platforms such as Materials Cloud, NOMAD, MaterialDigital, CAPeX and OPTIMADE, giving access to European materials data and digital tools through a single entry point. Its three focus areas are a common taxonomy, metadata standards and ontologies; simple access to data; and automated, interoperable workflows for experiments and simulations; it is also said to support Digital Product Passports [Reference 28].
The Communication set "by mid-2025" as a goal, and the project began in June 2026. That the target time and the project start differ is a fact as documented, and we make no judgement of it (commentary).
(5) NOMAD: a FAIR data-sharing infrastructure
NOMAD is a free, open-source service for managing and sharing materials-science data. The official website states that NOMAD is developed and operated by the FAIRmat consortium, that all published data are under CC-BY-4.0, that it extracts data from more than 60 file formats and that data can be published with a DOI. The site's display (checked 23 October 2026) shows 19,425,275 uploaded entries, 4,346,100 "represented materials" and 129.3 TB of files Sourced [Reference 9]. The displayed figures are current values of the site and change. We did not check how "entries" and "materials" are defined and counted.
CC-BY-4.0 is generally understood as a licence that permits copying, modification and commercial use on the condition that the source is credited (you need to check the licence text; this is not legal advice). A data infrastructure with one licence across all data, like NOMAD, has the advantage that the terms of use are clear. When you publish your own data too, it saves trouble later if you decide in-house, before publishing, under which licence, what you publish and what you do not (commentary).
5. Japan: strategy, the Materials DX Platform and NEDO
Japan's MI-related efforts overlap in time: a Cabinet Office strategy, MEXT's data-infrastructure programmes, projects of NEDO (under the Ministry of Economy, Trade and Industry) and projects of the Japan Science and Technology Agency (JST).
(1) The starting point: the stimulus of the US MGI, and the projects of JST, NIMS and NEDO
The 2019 know-how report of NIMS (National Institute for Materials Science) writes, as background, that the US MGI of June 2011 prompted the planning of materials informatics programmes in Japan at various levels, and that the Center for Research and Development Strategy (CRDS) of JST issued in August 2013 a strategic proposal recommending that materials informatics be made a strategic field of Japan Sourced [Reference 15]. Then, in fiscal 2015, the Materials research by Information Integration Initiative (MI2I), with NIMS as its hub, was selected as a project under JST's Innovation Hub Construction Support Program [References 15, 23]. According to the same report it was a short, five-year programme that drew on the management approach of the US Engineering Research Centers (ERC) [Reference 15].
The NEDO "Ultra-Advanced Materials Ultra-Fast Development Platform Technology" project (title in Japanese; our translation; known in Japan as the "Cho-Cho" project), implemented by NEDO under the Ministry of Economy, Trade and Industry, ran for seven years, fiscal 2016 to fiscal 2022. NEDO's public project record states a planned project scale of about 13.8 billion yen, and in the course of project costs the total NEDO contribution is 13.751 billion yen (the fiscal 2022 portion is a contracted amount). It was carried out by a technology research association joined by companies in materials and chemicals (16 companies in fiscal 2016, 18 in fiscal 2017), the National Institute of Advanced Industrial Science and Technology (AIST) and others. The final goal was to cut the number of trial products and the development time to 1/20 of conventional materials development, and the record's summary of results includes a statement in the process-technology item that the number was "reduced to almost 1/20" Sourced [Reference 14]. This statement of achievement is the project side's own summary, and we could not confirm whether an independent third party verified the 1/20. The record says it plans to spread the results after the project through a consortium for the use of data-driven materials design technology [Reference 14].
(2) Strategy: the Materials Innovation Strategy (decided 2021, revised 2025)
According to the Cabinet Office, the Materials Innovation Strategy (the "Materials Innovation Strength-Enhancement Strategy"; title in Japanese; our translation) was decided on 27 April 2021 by the Integrated Innovation Strategy Promotion Council and revised on 4 June 2025 Sourced [Reference 27]. The 2021 outline positions it as one of the important strategies, after AI, bio, quantum technology and the environment, aimed at strengthening the power to create materials innovation, and lists among the pillars of its action plan "promotion of data-driven R&D using materials data and manufacturing technology" (including the building of the Materials DX Platform) [Reference 10].
The 2025 revision sets out the "knowledge value chain", and as an action for "accelerating innovation" it lists further promotion of "Materials DX" centred on the use of data infrastructure (improving usability and accelerating data sharing, using AI and merging with robotics, and spreading data-driven R&D and producing results) [Reference 11]. The outline we checked has no amounts. In addition, the Cabinet Office page carries the expert panel's recommendation of 12 June 2026, "Measures for Implementing the Materials Innovation Strategy" (title in Japanese; our translation), but this is a recommendation of an expert panel, not a government decision. We did not check its content [Reference 27] Not yet confirmed.
(3) The Materials DX Platform: three programmes for creation, integrated management and use
MEXT is building a "Materials DX Platform" that advances research data in one piece: (1) creation (Advanced Research Infrastructure for Materials and Nanotechnology in Japan, ARIM), (2) integration and management (the NIMS data core hub, MDPF) and (3) use (the Data Creation and Utilization-Type Materials Research and Development Project, DxMT) Sourced [Reference 12].
- ARIM: started in fiscal 2021. At 26 universities and research institutes nationwide it provides shared use of about 1,200 instruments, and accumulates the data created from instrument use as structured data arranged to be usable for machine learning too. In the four years after the start it accumulated over 1 million records [Reference 12]. From fiscal 2025 a semiconductor platform (ARIM-SETI) has also been added [Reference 12].
- MDPF: NIMS serves as the hub for collecting, accumulating and using industry-academia-government materials data, and has provided the materials database (MatNavi), the data accumulation and sharing system (RDE), the repository of papers and data (MDR) and the integrated materials development system (MInt). From 23 December 2025 it provides "pinax", a playground for data analysis and machine learning, free of charge to researchers at universities and companies in Japan [Reference 13].
- DxMT: a programme that started after a one-year feasibility study, with five bases (Tohoku University, the University of Tokyo, Institute of Science Tokyo, Kyoto University and NIMS) working together. In the plan in MEXT's document, Phase I runs from September 2022 to March 2025, Phase II from April 2025 to March 2028 and Phase III from April 2028 to March 2031 [References 21, 22].
(4) ARIM data sharing: with conditions, and paid
According to MEXT's announcement of 26 August 2025, ARIM begins sharing the accumulated data starting with those for which the embargo (the period during which the creator of the data can use it exclusively for their own research; in principle up to two years from the end of the fiscal year of use) has passed. From 30 September 2025 it shares about 110,000 records (as of July 2025) under a prescribed fee. The form of provision is an annual subscription per single licence or group licence. The conditions for users are being a resident under the Foreign Exchange and Foreign Trade Act and belonging to a corporation established in Japan, applying for use and being granted permission. The data unify instrument-specific formats and consist of various experimental data including failed data, and are expected to grow by several hundred thousand records a year Sourced [Reference 12].
The same announcement says, as background, that "in the world, the United States and others have worked on (MI) as a national strategy from early on" [Reference 12]. That is, the Japanese government itself positions its effort with an eye on the US effort. We do not, however, evaluate the superiority of any country in this article.
There are two numbers in the same announcement: "over 1 million structured records accumulated in the four years after the start", and "about 110,000 shared from 30 September 2025 (as of July 2025)". The difference is because data are shared in order once the embargo (up to two years) has passed [Reference 12]. Not all accumulated data can be used from outside (commentary).
There are schemes like ARIM in which data measured on borrowed instruments are shared after a fixed period (the embargo). When you use a public shared instrument, check the handling of data (whose data they are, until when they are exclusive, and how they are shared afterwards) in the terms of use before you start. A mechanism that accumulates failed data is valuable as training data for MI (the ARIM announcement states that failed data are included) [Reference 12]. On the other hand, data that are confidential to your company require an arrangement that excludes them from sharing (commentary, not legal advice).
6. China: Materials Genome Engineering and "AI Plus"
China's MI-related efforts are called "Materials Genome Engineering". This section summarizes statements of official Chinese-language documents in English and writes only what we could confirm.
(1) The National Key R&D Program special project on Materials Genome Engineering
The 2017 application guideline of the Ministry of Science and Technology (Annex 12; in Chinese) says that the special project on key technologies and a support platform for Materials Genome Engineering (our translation of the project name) is a special project of the National Key R&D Program for carrying out tasks set out in the State Council's "Made in China 2025" and the 13th Five-Year National Science and Technology Innovation Plan. Its overall aim is to halve the R&D cycle and halve the cost of new materials, merging high-throughput calculation (theory), high-throughput experiments (preparation and characterization) and dedicated databases, and changing the way of developing new materials from the conventional "experience guides experiment" to a new "theory predicts, experiment verifies" Sourced [Reference 16].
The main contents are to build three demonstration platforms (high-throughput calculation, high-throughput preparation and characterization, and dedicated databases); to develop four key technologies (multiscale high-throughput calculation methods and software, high-throughput material preparation, high-throughput characterization and evaluation of service behaviour, and materials big-data technology); and to demonstrate them on energy materials, biomedical materials, rare-earth functional materials, catalytic materials, special alloys and others. The special project has 40 key research tasks and an implementation period of five years (2016 to 2020); 13 tasks were started in 2016, 16 tasks were planned for 2017, supporting 18 to 36 projects, with a total state-funding estimate of 324 million yuan, and enterprise-led projects must provide their own matching funds of at least the same amount as the state funding Sourced [Reference 16].
How to read it: the 324 million yuan is an "estimate" in the 2017 guideline (a planning figure), not the total for the whole period of the special project and not an actual allocation or expenditure. We could not confirm the results of the special project, any successor after 2020, the amount enacted or the spending Not yet confirmed.
(2) The State Council opinion on the "AI Plus" action (2025)
In August 2025 (dated 21 August and published on 26 August) the State Council issued the "Opinion on Deepening the Implementation of the 'AI Plus' Action" (Guofa [2025] No. 11; title in Chinese; our translation). The first of the six key actions of "AI Plus" is "AI plus science and technology", whose section on "accelerating scientific discovery" lists exploring new AI-driven research paradigms, building and applying scientific foundation models, upgrading basic research platforms and major science and technology infrastructure to be intelligent, and building open, shared, high-quality scientific datasets Sourced [Reference 17]. In the text we retrieved, we could not confirm a passage naming "materials". We do not treat this document as a materials MI strategy (read it as a policy on using AI in science in general).
The 2017 guideline lists "dedicated databases" as one of the demonstration platforms [Reference 16]. But we could not confirm from official sources to whom and on what terms materials databases are open today. We do not write estimated amounts or scales that appear in private surveys or press reports, because their sources and scope are not fixed.
7. A timeline, and why not to compare
We line up the documents and events so far on a timeline. The positions of the points are approximate.
| Country / region | Nature of the main documents | Amounts and targets stated (original currency, wording of the document) | What we could not confirm |
|---|---|---|---|
| United States | Multi-agency strategic plan, executive order, DOE call [1 to 4] | MGI: twice as fast, at a fraction of the cost (target). Genesis: the goal of doubling the productivity and impact of R&D within a decade (target). DOE call: US$293 million (call), Phase I US$0.5-0.75 million, Phase II US$6-15 million | Amounts for MGI. The budget of the Genesis Mission as a whole. Results of selection in the call |
| Europe | Research framework, policy communication, legislative proposal (planned), grant agreement [5 to 8] | Horizon Europe 2021-2027: indicative EUR 93.5 billion. Next framework 2028-2034: EUR 175 billion (proposal). MaterialsCommons: EUR 27,998,554.91 (grant agreement) | The proposal and content of the Advanced Materials Act. Enactment of the next framework |
| Japan | Cabinet Office strategy (decision), expert recommendation, MEXT programmes, NEDO project record [10 to 15, 21, 27] | NEDO project: about 13.8 billion yen (plan), total NEDO contribution 13.751 billion yen. Target: trial products and development time to 1/20 | Budget amounts for ARIM and DxMT. The content of the June 2026 recommendation. Independent verification of the 1/20 |
| China | Application guideline of the Ministry of Science and Technology, State Council opinion [16, 17] | Materials Genome Engineering: halve the R&D cycle and cost (target). 2017 state-funding estimate: 324 million yuan (estimate) | Results, enacted amount and spending of the special project. Terms of access to materials data infrastructure |
All Sourced (official documents, checked 23 October 2026). The arrangement of the table is Our calculation (organization). Amounts in different rows differ in scope, period, stage and currency, so they cannot be compared or added.
Four things to check before comparing
- Stage: proposal, estimate, framework, call, contract or actual (Fig. 2).
- Scope: MI only, materials as a whole, or science as a whole (for example, China's "AI Plus" covers science generally). The same "X billion" has a different breadth of target.
- Period and fiscal year: one year, five or seven? The division of fiscal years differs by country.
- Currency and nature of the document: keep the original currency. Law, order, strategy, call, guideline or recommendation: the legal effect differs.
Official documents contain comparisons between countries too. For example, the Commission's Communication COM(2024) 98 states that industrial R&I investment in advanced materials (2020) was EUR 19.8 billion in the EU, which is not even half of the EUR 50.3 billion of the United States, with South Korea at EUR 19.6 billion, Japan at EUR 14.0 billion and Chinese industry at EUR 7.7 billion [Reference 6]. This is the Commission's claim, and the figures depend on the definitions of the source (what counts as "advanced materials" and as industry investment). We do not use these figures to rank countries.
8. Terms of access to data infrastructure: a materials engineer's view
For a materials engineer, the "data infrastructure" of a national strategy is a practical question: "where is the materials data usable for our own MI, and on what terms can it be used?" We line up the infrastructures confirmed in this article by terms of use (checked 23 October 2026).
| Infrastructure | Operator (country / region) | Size of data (as displayed on the site) | Terms of use (official statements) | What we could not confirm |
|---|---|---|---|---|
| NOMAD [9] | FAIRmat (Europe) | 19,425,275 entries, 4,346,100 materials, 129.3 TB of files | All data under CC-BY-4.0. Free and open source | The definitions of how "entries" and "materials" are counted |
| NIMS MDR and pinax [13, 18] | NIMS (Japan) | No figure displayed (in what we checked) | MDR: the licence stated on each dataset page. pinax: free for researchers at universities and companies in Japan (from 23 December 2025) | The number of datasets. Individual licences |
| ARIM data sharing [12] | MEXT programme (Japan) | Over 1 million accumulated; about 110,000 shared (as of July 2025) | Paid annual subscription. Residents under the Foreign Exchange Act affiliated with a corporation in Japan. Embargo of up to two years | The amount of the fee. Details of whether foreign researchers and companies may use it |
| JARVIS (NIST) [19] | NIST (United States) | Over 80,000 materials, over 1 million properties, over 150,000 users | Free registration is required to use the database. Funding: MGI and CHIPS | Licences of individual data |
| MaterialsCommons [8, 28] | EU-funded project (Europe) | Before launch (started June 2026) | A plan to link existing infrastructures and make them a single entry point | The timing and terms of release |
| Materials Project | A DOE effort [description in Reference 2] | Could not be confirmed | Could not be confirmed (the site shows a verification screen) | Terms of use and data size |
All Sourced (statements on each official site, checked 23 October 2026). Figures on the sites are current values and change. The arrangement of the table is Our calculation (organization). Interpretation of licence texts depends on each infrastructure's text (this is not legal advice).
As for the Materials Project, beyond the context of MGI [Reference 2], which describes it as a long-term DOE effort that pre-computes properties of materials and makes them public, we could not confirm the terms of use of the site. Because the site displayed a verification screen (bot protection), we did not go past it.
Public data infrastructures are updated and change release. When you use them, add to your own records the name of the infrastructure and the date (release) you obtained the data, whether the data are calculated or measured (the "three columns" of our explainer on applications to semiconductor and electronic materials), and the licence and terms of use. If you mix data under different licences and publish or distribute them, you need to check whether the combination is allowed (commentary, not legal advice).
When you write "the state invests X billion" or "to be realized by year Y" in a business plan or proposal, add the stage of the amount (proposal, estimate, call, contract), the source document and its date, and the scope of the target (MI only or materials as a whole). For example, DOE's US$293 million is "the amount of a call", Europe's EUR 175 billion is "a proposal", and NEDO's 1/20 is "a statement of achievement organized by the project side". Without the stage, readers mistake a plan for a result (commentary).
9. Limits, what we could not confirm / summary
(1) Amounts
We have not written amounts for the US MGI, the budget of the Genesis Mission as a whole, the results of selection in DOE's call, the budgets of Japan's ARIM and DxMT, or the results and enacted and spent amounts of China's special project, because we could not confirm them. The amounts we could confirm are written with their stage.
(2) Recent developments
A proposal for the European Advanced Materials Act, enactment of the next European research framework, the content of the Japanese expert recommendation (June 2026) and how far the deadlines of the Genesis Mission executive order have been met may change after the checking date. This article does not treat them as settled systems.
(3) Chinese-language documents
The guideline of the Ministry of Science and Technology and the opinion of the State Council are Chinese official documents summarized in English. We tried to avoid errors in the summary, but please check the exact wording in the original.
(4) Countries and topics not covered
We do not cover countries such as the United Kingdom, South Korea and India. Efforts of MI-related companies, universities and institutes in each country, systems of export control and intellectual property, and the treatment of personal information and security aspects of materials data are outside the scope of this article.
(5) The contents of data infrastructure
For some items, such as the quality and size definitions of each data infrastructure, individual licences and the terms of use of the Materials Project, we could not confirm them. Figures on sites are as of the checking date.
- The nature of national MI strategy documents differs by country. An executive order, a policy communication, a government decision, an application guideline and an expert recommendation are different things Sourced.
- Amounts come in stages. DOE's US$293 million is a call, Europe's next EUR 175 billion is a proposal, NEDO's about 13.8 billion yen is a plan, and NEDO's total contribution of 13.751 billion yen is close to an actual. Europe's MaterialsCommons is a grant agreement of EUR 27,998,554.91 Sourced.
- A target is different from achievement. The "twice as fast" of MGI, the "1/20" of NEDO and the "halving of cycle and cost" of China are targets or statements by the project side, and we could not find independent verification Not yet confirmed.
- Terms of access to data infrastructure differ by infrastructure. NOMAD is CC-BY-4.0, ARIM is paid with eligibility limited to residents and corporations in Japan, and MDR has a licence per dataset Sourced.
- We do not rank countries. Numbers that differ in scope, period, stage and currency cannot be compared or added.
10. How to read the yardsticks and standards
Unlike yardsticks for the performance of a technology, there are almost no standards common to countries for gauging the "progress" of a national MI strategy. The yardsticks usable in this article are four: the stage of money, the difference between a target and achievement, how FAIR the data are and the terms of access, and evaluation of MI results. Read each with awareness of what it measures and what it does not. We organize what we could confirm by kind (checking date: 23 October 2026).
| What is evaluated | Representative evaluation method or source (name, reference) | Standards and guidance (name, edition, source checked) | Cautions when reading |
|---|---|---|---|
| Size and stage of money | EC's Horizon Europe page (distinguishing indicative and proposed) [5], CORDIS (grant agreement) [8], NEDO project record (plan and contribution) [14], DOE announcement (call) [4], guideline of the Ministry of Science and Technology (estimate) [16] Sourced | Could not be confirmed (an accounting standard common to countries) | Check the stage (proposal, estimate, framework, call, contract, actual). Do not convert exchange rates, add or compare |
| Targets and achievement | MGI's target (twice as fast) [1], NEDO's target (trial products and development time to 1/20) and the project side's statement of achievement [14], China's target (halve the cycle and cost) [16] Sourced Not yet confirmed | Could not be confirmed (a standard for measuring achievement of targets in a common way) | The definition of the target (what, compared with what) differs by country. Independent verification of achievement is unconfirmed. A statement of achievement is the project side's own summary |
| How FAIR the data are | The MGI plan (FAIR as a foothold for "AI-ready" data) [2], the EU Communication (the materials commons is based on FAIR) [6] Sourced | FAIR principles (Wilkinson et al., 2016) [24] / ISO/IEC 5259-1:2024 (data quality for AI, Part 1: overview, terminology and examples, July 2024) [25] Sourced | FAIR is a set of principles (recommended) and not legally binding. The full text of ISO/IEC is paid and the requirements were not checked |
| Terms of access to data | NOMAD: CC-BY-4.0 [9] / MDR: licence per dataset [18] / ARIM: paid annual subscription and eligibility for residents and corporations in Japan [12] Sourced | Could not be confirmed (an international standard defining how materials data are released) | A licence is a condition of use, not a standard. Conditions, fees and eligible users differ by infrastructure |
| Evaluation of MI results | Matbench, Matbench Discovery and JARVIS-Leaderboard (see our explainer on how to read evaluation and reproducibility in MI) | Could not be confirmed (an international standard specific to evaluating MI). AI terms: ISO/IEC 22989:2022 (July 2022) [26] Sourced | We could not confirm which benchmark the documents of national strategies use to evaluate results |
| Status of the document (binding force) | Executive order ("consistent with applicable law and subject to available appropriations") [3] / the distinction between a Communication (policy) and a legislative proposal (the Advanced Materials Act) [6, 7] / the distinction between a decision of a strategy and an expert recommendation [27] Sourced | - (depends on the text of each document and the responsible body) | Interpretation of legal effect depends on the text of each document and the responsible body. This article is not legal advice |
The arrangement of the table is Our calculation (organization). Names, dates and status were checked in the official documents of each body (checked 23 October 2026). The full texts of standards are paid, so we stay with scope statements.
Cautions when reading (for materials engineers)
- Tell apart "announcement", "plan", "proposal" and "contract": news reports them in the same tone. Money moves after a budget is enacted and a contract is signed (commentary).
- A "target" is no guarantee of achievement: the "twice as fast" of MGI, the "1/20" of NEDO and the "half" of China are targets. Even when a project side organizes it as "achieved", check whether there is independent verification.
- "Existing" and "usable" are different for a data infrastructure: check the terms of release, fees and eligible users. Some infrastructures have eligibility conditions, such as being limited to residents and corporations in the country.
- Dates and status are updated: the European Advanced Materials Act, the next research framework, the Japanese expert recommendation and the progress of the US Genesis Mission may change after the checking date. When you quote, add the checking date.
- Do not make a league table: this article does not rank countries. Be wary of articles that line up amounts without the four checks of chapter 7.
11. Glossary
- MGI (Materials Genome Initiative)
- A US federal effort begun in 2011 to speed up materials development with data and computation.
- MII (Materials Innovation Infrastructure)
- The collection of computational, experimental and data tools described in the MGI plan.
- FAIR principles
- Principles for making data findable, accessible, interoperable and reusable.
- AI-ready data
- Data arranged in a form that machine learning can use.
- Executive order
- An order issued by the US President to executive agencies. Amounts are decided separately by Congress.
- Appropriation
- An amount that Congress has decided as a budget to be actually spent.
- RFA / NOFO
- A call for applications. It states the conditions of application, the range of amounts and deadlines.
- Genesis Mission
- A national effort begun by a US executive order in November 2025 to accelerate scientific discovery with AI.
- Horizon Europe
- The EU's main research and innovation funding programme.
- Communication
- A document in which the European Commission sets out policy. Separate from a legislative proposal.
- Grant agreement
- A contract that fixes the amount and period of an EU grant to a project.
- CORDIS
- The EU site that publishes information on research projects.
- Materials commons
- A lasting digital infrastructure for advanced materials research and innovation envisaged by the EU.
- NOMAD / FAIRmat
- A service for managing and sharing materials-science data / the consortium that develops and operates it.
- CC-BY-4.0
- A kind of licence that allows broad use on the condition that the source is credited.
- ARIM
- A programme of the Japanese MEXT in which 26 institutions share advanced instruments and accumulate the data they produce.
- MDPF / MDR
- The NIMS data core hub / the repository publishing papers and data on materials.
- DxMT
- The data creation and utilization materials R&D project. Five bases advance research that uses data.
- Embargo (non-sharing period)
- A period during which only the data creator can use the data; sharing follows afterwards.
- Structured data
- Data arranged in a machine-friendly form (csv, tables and so on).
- Materials Genome Engineering
- The Chinese name of the effort to speed up materials development with high-throughput calculation, experiment and data.
- Special project / estimate
- A project in a specific field of the National Key R&D Program / an estimated amount written in a plan.
- Expert-panel recommendation
- A proposal of a panel of experts. Separate from a government decision.
12. References (primary sources)
- Materials Genome Initiative official website (statement of the goal, the 2024 Challenges and others; checked 23 October 2026) - mgi.gov
- NSTC Subcommittee on the Materials Genome Initiative "Materials Genome Initiative Strategic Plan", November 2021 (full text checked) - mgi.gov/media/256
- The White House "Launching the Genesis Mission", Executive Order 14363, 24 November 2025 (full text checked) - whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/
- U.S. Department of Energy "Energy Department Announces $293 Million in Funding to Support Genesis Mission National Science and Technology Challenges" (March 2026; full text checked) - energy.gov/articles/...-genesis-mission-national-science
- European Commission "Horizon Europe" (research and innovation page; indicative funding amount and proposal for the next framework; checked 23 October 2026) - research-and-innovation.ec.europa.eu/.../horizon-europe_en
- European Commission "Advanced Materials for Industrial Leadership", COM(2024) 98 final, 27 February 2024 (full text checked) - research-and-innovation.ec.europa.eu/document/download/0fcf06ea-..._en
- European Commission "Commission seeks feedback on the future Advanced Materials Act" (research and innovation news, 21 October 2025) - research-and-innovation.ec.europa.eu/news/...-2025-10-21_en
- CORDIS "Federated Digital Infrastructure for European Advanced Materials Research and Development (MaterialsCommons)", Grant agreement ID 101293545 - cordis.europa.eu/project/id/101293545
- NOMAD official website (development and operation by FAIRmat, CC-BY-4.0, displayed counts; checked 23 October 2026) - nomad-lab.eu/nomad-lab/
- Cabinet Office (Japan) "Materials Innovation Strategy (outline)", April 2021 (in Japanese) - www8.cao.go.jp/cstp/material/material_gaiyo.pdf
- Cabinet Office (Japan) "Materials Innovation Strategy: through building a knowledge value chain (outline)", 4 June 2025 (in Japanese) - www8.cao.go.jp/cstp/material/material_2025_gaiyo.pdf
- MEXT (Japan) "Building and starting shared use of a large-scale materials data infrastructure", 26 August 2025 (in Japanese; full text checked) - mext.go.jp/b_menu/houdou/mext_01556.html
- NIMS "Launch of the data analysis and machine learning playground 'pinax'", 23 December 2025 (in Japanese) - nims.go.jp/press/2025/12/.../202512230.pdf
- NEDO Project record (public version, material for the post-project evaluation) of the "Ultra-Advanced Materials Ultra-Fast Development Platform Technology" project (title in Japanese; our translation) - nedo.go.jp/content/100947916.pdf
- NIMS (Ito S.) "Innovation Hub Construction Support Program: know-how report, Materials research by Information Integration Initiative", September 2019 (in Japanese; full text checked) - jst.go.jp/ihub/files/seika-nims.pdf
- Ministry of Science and Technology (China) Guideline for applications in 2017 for the special project on key technologies and a support platform for Materials Genome Engineering (our translation of the title), Annex 12 (in Chinese; original text checked) - service.most.gov.cn/u/cms/static/201610/141014524nry.pdf
- State Council (China) Opinion on deepening the implementation of the "AI Plus" action, Guofa [2025] No. 11, August 2025 (our translation of the title; in Chinese; original text checked) - gov.cn/zhengce/content/202508/content_7037861.htm
- NIMS Materials Data Repository (MDR) - mdr.nims.go.jp
- NIST JARVIS (Joint Automated Repository for Various Integrated Simulations) official website, checked 23 October 2026 - jarvis.nist.gov
- U.S. Department of Energy Genesis Mission official website, checked 23 October 2026 - genesis.energy.gov
- MEXT (Japan) "Overview of the Data Creation and Utilization-Type Materials Research and Development Project (DxMT)", 31 July 2024 (in Japanese) - mext.go.jp/content/20241028_mxt-nanozai01_000038609-3.pdf
- NIMS Page of the DxMT data-collaboration subcommittee secretariat (in Japanese) - nims.go.jp/aesd/unit/dxmt.html
- JST Innovation Hub Construction Support Program: outline of selected projects (Materials research by Information Integration Initiative; in Japanese) - jst.go.jp/ihub/gaiyo.html
- Wilkinson M. D. et al. "The FAIR Guiding Principles for scientific data management and stewardship", Scientific Data 3, 160018 (2016), DOI 10.1038/sdata.2016.18 - doi.org/10.1038/sdata.2016.18
- ISO/IEC "ISO/IEC 5259-1:2024 Artificial intelligence - Data quality for analytics and machine learning (ML) - Part 1: Overview, terminology, and examples" (ISO page) - iso.org/standard/81088.html
- ISO/IEC "ISO/IEC 22989:2022 Information technology - Artificial intelligence - Artificial intelligence concepts and terminology" (ISO page) - iso.org/standard/74296.html
- Cabinet Office (Japan) "Materials strategy" page (dates of decisions and the list of expert-panel recommendations; in Japanese; checked 23 October 2026) - www8.cao.go.jp/cstp/material/material.html
- IMDEA Materials Institute "MaterialsCommons to transform how Europe develops advanced materials" (announcement by a participating institution, 2 July 2026) - materials.imdea.org/materialscommons-to-transform-how-europe-develops-advanced-materials/
13. Claim-to-source audit
| Claim in the text | Basis | Label |
|---|---|---|
| MGI described as a federal multi-agency initiative to discover, manufacture and deploy advanced materials twice as fast and at a fraction of the cost; the 2024 Challenges | MGI official website Reference 1 https://www.mgi.gov/ | Sourced |
| MGI launched in 2011 and the 2021 plan updating the December 2014 plan; three goals, FAIR and AI-ready data, NSF MIP (2016); no amounts in what we checked | MGI Strategic Plan (November 2021) Reference 2 https://www.mgi.gov/media/256 | Sourced |
| Executive Order 14363 (24 November 2025): roles of DOE and the APST, the Platform, deadlines of 60/90/120/240/270 days, domains of challenges (including critical materials), "subject to available appropriations" | Text of the executive order Reference 3 https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/ | Sourced |
| DOE's US$293 million call (DE-FOA-0003612), Phase I US$0.5-0.75 million for nine months, Phase II US$6-15 million for three years, deadlines (28 April and 19 May 2026) | DOE announcement Reference 4 https://www.energy.gov/articles/energy-department-announces-293-million-funding-support-genesis-mission-national-science | Sourced |
| Horizon Europe 2021-2027 indicative amount of EUR 93.5 billion; the proposal of EUR 175 billion for 2028-2034 (16 July 2025) and negotiations | European Commission page (checked 23 October 2026) Reference 5 https://research-and-innovation.ec.europa.eu/funding/funding-opportunities/funding-programmes-and-open-calls/horizon-europe_en | Sourced Not yet confirmed |
| COM(2024) 98 (27 February 2024): 10 to 30 years of development, the example of AI predicting nearly 400,000 stable crystal structures, the materials commons by mid-2025, based on FAIR, comparison of industrial investment (2020) | Text of the Communication Reference 6 https://research-and-innovation.ec.europa.eu/document/download/0fcf06ea-c242-44a6-b2cb-daed39584996_en | Sourced |
| Call for evidence on the Advanced Materials Act (opened 21 October 2025, until 13 January 2026), expected to be proposed in 2026, building on the Communication | European Commission news Reference 7 https://research-and-innovation.ec.europa.eu/news/all-research-and-innovation-news/commission-seeks-feedback-future-advanced-materials-act-2025-10-21_en | Sourced Not yet confirmed |
| MaterialsCommons: grant agreement 101293545, EU signature 12 May 2026, 1 June 2026 to 31 May 2030, total cost and EU contribution both EUR 27,998,554.91 | CORDIS Reference 8 https://cordis.europa.eu/project/id/101293545 | Sourced |
| NOMAD: developed and operated by FAIRmat, all data under CC-BY-4.0, 19,425,275 entries, 4,346,100 materials and 129.3 TB (displayed on the checking date) | NOMAD official website Reference 9 https://nomad-lab.eu/nomad-lab/ | Sourced |
| The position of the 2021 Materials Innovation Strategy and its pillar of promoting data-driven R&D | Cabinet Office outline (April 2021) Reference 10 https://www8.cao.go.jp/cstp/material/material_gaiyo.pdf | Sourced |
| The revision of 4 June 2025: the knowledge value chain, further promotion of Materials DX (accelerating data sharing, AI and robotics, data-driven R&D) | Cabinet Office outline (4 June 2025) Reference 11 https://www8.cao.go.jp/cstp/material/material_2025_gaiyo.pdf | Sourced |
| The three programmes of the Materials DX Platform (ARIM, MDPF, DxMT); ARIM (started fiscal 2021, 26 institutions, about 1,200 instruments, over 1 million records); data sharing (30 September 2025, about 110,000 records, paid annual subscription, user conditions, embargo up to two years, including failed data); the revision of the strategy (4 June 2025) | MEXT announcement (26 August 2025) Reference 12 https://www.mext.go.jp/b_menu/houdou/mext_01556.html | Sourced |
| MDPF services (MatNavi, RDE, MDR, MInt) and free provision of pinax (from 23 December 2025, researchers at universities and companies in Japan) | NIMS announcement Reference 13 https://www.nims.go.jp/press/2025/12/a5kbo700000092mp-att/202512230.pdf | Sourced |
| NEDO project: seven years (fiscal 2016-2022), planned scale of about 13.8 billion yen, total NEDO contribution of 13.751 billion yen (fiscal 2022 contracted), final goal of 1/20 trial products and development time, the statement of "almost 1/20" in process technology | NEDO project record Reference 14 https://www.nedo.go.jp/content/100947916.pdf | Sourced |
| CRDS's proposal of August 2013, the stimulus of the MGI (June 2011), the selection of MI2I in fiscal 2015 (five years, drawing on the ERC) | NIMS know-how report (September 2019) Reference 15 https://www.jst.go.jp/ihub/files/seika-nims.pdf | Sourced |
| China's special project: overall aim (halve cycle and cost), three demonstration platforms, four key technologies, 40 tasks and five years (2016-2020), 2017 estimate of 324 million yuan, 18 to 36 projects, matching funds of at least 1:1 | 2017 guideline of the Ministry of Science and Technology (in Chinese) Reference 16 https://service.most.gov.cn/u/cms/static/201610/141014524nry.pdf | Sourced |
| The State Council opinion on "AI Plus" (Guofa [2025] No. 11, dated 21 August 2025, published 26 August) and the contents of the "AI plus science and technology" section | State Council document (in Chinese) Reference 17 https://www.gov.cn/zhengce/content/202508/content_7037861.htm | Sourced |
| MDR's "licence stated on each dataset page" | NIMS MDR Reference 18 https://mdr.nims.go.jp/ | Sourced |
| JARVIS: established January 2017, funding from MGI and CHIPS, over 80,000 materials, over 1 million properties, registration required | JARVIS official website (checked) Reference 19 https://jarvis.nist.gov/ | Sourced |
| The goal of the Genesis Mission (doubling productivity and impact within a decade) and the examples for materials | DOE Genesis website Reference 20 https://genesis.energy.gov/ | Sourced |
| DxMT Phase I (September 2022 to March 2025), Phase II (April 2025 to March 2028), Phase III (April 2028 to March 2031) | MEXT overview (31 July 2024) Reference 21 https://www.mext.go.jp/content/20241028_mxt-nanozai01_000038609-3.pdf | Sourced |
| DxMT started after a one-year feasibility study, with five bases (Tohoku University, the University of Tokyo, Institute of Science Tokyo, Kyoto University and NIMS) working together | NIMS secretariat page Reference 22 https://www.nims.go.jp/aesd/unit/dxmt.html | Sourced |
| MI2I as a project of JST's Innovation Hub Construction Support Program, with NIMS as the hub | JST page Reference 23 https://www.jst.go.jp/ihub/gaiyo.html | Sourced |
| The FAIR principles are a concise and measurable set of principles for improving the reusability of data | Abstract of Wilkinson et al. (2016) Reference 24 https://doi.org/10.1038/sdata.2016.18 | Sourced |
| Numbers, titles and publication dates of ISO/IEC 5259-1:2024 (July 2024) and ISO/IEC 22989:2022 (July 2022) | ISO pages Reference 25 https://www.iso.org/standard/81088.html Reference 26 https://www.iso.org/standard/74296.html | Sourced |
| Decision dates of the Materials Innovation Strategy (27 April 2021 and 4 June 2025) and the listing of the expert-panel recommendation of 12 June 2026 | Cabinet Office page (checked 23 October 2026) Reference 27 https://www8.cao.go.jp/cstp/material/material.html | Sourced |
| The MaterialsCommons consortium (26 academic partners, 21 associated partners, 12 associated collaborators), the coordinating institute, the kick-off (17 June 2026), the aim and three focus areas | Announcement by IMDEA (2 July 2026) Reference 28 https://materials.imdea.org/materialscommons-to-transform-how-europe-develops-advanced-materials/ | Sourced |
| That amounts for the US MGI, the budget of the Genesis Mission as a whole and the results of selection in DOE's call, budgets of ARIM and DxMT, the content of the June 2026 expert recommendation, independent verification of NEDO's 1/20, the results and enacted amount of China's special project, the terms of access to Chinese materials data infrastructure and the terms of use of the Materials Project could not be confirmed and are not stated | Not found in the primary sources we could reach in this survey | Commentary |
| The organization into five stages of money, the table of document types, the four checks, and readings such as "'existing' and 'usable' differ for data infrastructure" and "the size of a budget is not a yardstick of progress", and the materials engineer's views | Our own organization and commentary based on published content; not views stated by the institutions concerned | Commentary |
| That figures 2 to 8 are explanatory organization, not observations of real data, that figures 1 and 9 are frames for AI-generated images, and that countries are not ranked | Note by this article | Commentary |
Last updated: 23 October 2026 / Sources are limited to primary ones (official documents of governments, the European Commission, MEXT, the Cabinet Office, NEDO, the Ministry of Science and Technology and the State Council of China; CORDIS; official websites of NOMAD, NIMS and NIST). We do not write amounts or situations we could not confirm. Chinese-language documents are summarized in English; please check the exact wording in the original. This article does not rank countries and does not give advice on investment, policy or law. Amounts for the US MGI, the budget of the Genesis Mission as a whole and the results of selection, budgets of ARIM and DxMT, the content of the Japanese expert recommendation (June 2026), independent verification of NEDO's 1/20, the results and enacted amount of China's special project, and the terms of use of the Materials Project could not be confirmed and are therefore not stated. All figures are explanatory concept diagrams; Figs. 1 and 9 are ImageGen-generated conceptual illustrations.