INSTITUTION PROFILE / MATERIALS INFORMATICS
NEDO
The national agency that carries out Japan's materials informatics R&D
NEDO is a national research and development agency overseen by Japan's Ministry of Economy, Trade and Industry (METI). It does not do materials research itself. Its job is to use the government's R&D budget to run open calls, then manage and evaluate the projects of the companies, universities and research institutes it selects. In materials informatics (MI, which means using AI and data to speed up materials development), the best-known example is the "Ultra-Ultra project" of FY2016 to FY2022. This profile looks at one question: how far NEDO's influence over MI reaches, and what it cannot decide.
- NEDO in 30 seconds
- What kind of body it is: the four types of influence
- What it does for MI
- How far its influence reaches (powers and geography)
- Who is involved and how projects are organised
- Outlook
- Glossary, references and claim-to-source audit
1. NEDO in 30 seconds
NEDO is the body that carries out Japan's MI research and development; it is not the one that sets the direction. Its biggest MI effort was the Ultra-High-Throughput Design and Prototyping Technology for Ultra-Advanced Materials Development Project (the "Ultra-Ultra project", FY2016 to FY2022). For organic functional materials, it aimed to cut the number of trial batches and the development time to one-twentieth of the conventional level. An outside expert panel assessed the goal as "largely achieved", while noting that the range of materials it applies to is limited Sourced. As of 22 October 2026, NEDO's list of FY2026 calls for proposals in "materials and manufacturing" shows no calls at all Sourced. For the next large MI project, METI has requested a new ¥7.0 billion for FY2027, to be carried out through NEDO, but this is not yet approved Not yet confirmed.
2. What kind of body it is: the four types of influence
Unlike the companies in this series, the institutions that appear in MI coverage do not develop materials themselves. What separates them is the route by which they influence the field, and that gives four types. NEDO belongs to the "funding" type.
NEDO's strength is the practical work of turning public budgets into actual projects. The broad direction is set by the state beforehand. For the Ultra-Ultra project, for example, NEDO's evaluation materials show that the project was placed within the national Science and Technology Basic Plan and a technology strategy for functional materials Sourced.
3. What it does for MI
The NEDO projects with an MI component that we could confirm from published documents are the following Sourced.
| Project | Period and budget | Content | Main performers |
|---|---|---|---|
| Ultra-High-Throughput Design and Prototyping Technology for Ultra-Advanced Materials Development Project (the "Ultra-Ultra project") | FY2016 to FY2022 Total cost ¥13.8 billion (an estimate for seven years) | For organic functional materials, combine computational science (simulation), fast prototyping processes and advanced measurement to generate data and train AI on it. The goal was one-twentieth of the usual trial batches and development time. From FY2019 it also added an AI tool that automatically extracts property data from polymer papers | AIST and a research association of 18 companies worked together at a central laboratory in Tsukuba. Ten universities and other bodies took part (as of March 2022) |
| Project to build a materials development and evaluation system for aircraft engines | FY2021 to FY2025 FY2025 budget ¥700 million | One work item develops a way to use MI to search for alloys with the needed properties. The project also covers more advanced manufacturing processes and a materials database that can stand up to certification | Handled by NEDO's Aerospace Department. We did not check the list of performers |
| MI for fuel cell and water electrolysis materials (now part of the project named "R&D to strengthen common foundations for expanding hydrogen use") | NIMS's part ran from August 2020 to March 2025 (planned). Still reported at a July 2026 conference | A database and web app that collect fuel cell experiment data in a common format; 28,850 files had accumulated by 13 April 2026. One reported case tested in experiments a molecule proposed by AI | NIMS, Hokkaido University, Kyoto University, Nagoya University, Toyota Technological Institute and Waseda University |
| Survey on MI for hydrogen production by water electrolysis | FY2022 | A survey of how MI could help develop catalysts and membranes for electrolysis. The call drew one proposal; the planned performer was Hokkaido University | Hokkaido University |
"Development time cut to one-twentieth" does not mean every material can now be made twenty times faster. The project's materials report that, for five model materials (organic functional materials chosen for their end products), the number of trial batches and the development time fell to roughly one-twentieth or less Sourced. That is NEDO's own account to its evaluation committee. The outside experts' overall assessment accepts that the goal was "largely achieved", but says the range of application is "limited" and that the data platforms are "not comprehensive databases" Sourced. For a materials maker, this means you need to check case by case whether it carries over to your own material system.
4. How far its influence reaches (powers and geography)
| Aspect | What NEDO has | What NEDO does not have |
|---|---|---|
| Powers | It runs open calls, decides who carries out the work, manages progress and submits projects to evaluation by outside experts. It has data infrastructure built within projects (the Ultra-Ultra project set up five data platforms) | The power to set MI policy or strategy (that sits with METI, the Cabinet Office and others). The power to set international standards. The power to approve budgets |
| Geography | Projects funded by Japan's budget. Awardees are mainly national institutes, universities and companies in Japan | Direct power over foreign companies or other countries' policies |
| Effect on companies | It gives participating companies shared core technology, data and a place for evaluation. In the Ultra-Ultra project, 18 companies took part as members of the research association | The power to require non-participating companies to use the results |
| Size | 1,640 staff. A ¥505.2 billion budget plus ¥11.37 trillion in funds (all fields combined) | The total spent on MI alone is not published. The ¥13.8 billion for the Ultra-Ultra project is a seven-year estimate |
The fund budgets of more than ¥11 trillion cover all of NEDO's fields, and the semiconductor and post-5G funds make up most of it (post-5G ¥3.59 trillion, Green Innovation ¥2.76 trillion, specified semiconductors ¥2.17 trillion, and others) Sourced. The amount used for MI is far smaller, and the total for MI alone is not published Not yet confirmed. The "¥13.8 billion" for the Ultra-Ultra project is an estimated total for seven years, not actual spending.
5. Who is involved and how projects are organised
In the Ultra-Ultra project, AIST and the Research Association of High-Throughput Design and Development for Advanced Functional Materials (ADMAT) worked together at a central laboratory in Tsukuba, under a project leader from AIST Sourced. The association's 18 member companies include Konica Minolta, Tosoh, Toray and Zeon. In a call for proposals in FY2019, two national institutes, three universities and six companies were selected as performers, and the structure chart also names Mitsui Chemicals, a company profiled in this series (we could not tell from the chart which role it had) Sourced. As an example of results, in June 2022 NEDO announced, together with AIST, ADMAT and Zeon, a technique that combines several AI models to predict the properties of complex materials. The announcement says it can make the predictions in one twenty-thousandth of the time or less for material systems that earlier AI could not handle Sourced. That is a statement by NEDO and the participating bodies, and whether it carries over to other material systems has not been confirmed Not yet confirmed.
6. Outlook
| What can be read | Primary source behind it | Category |
|---|---|---|
| If METI's new FY2027 project (¥7.0 billion) is approved, NEDO would carry it out. Approval is not settled | METI's request materials describe a structure of the state, then NEDO, then national institutes, universities and companies. A budget request is made before approval | Not yet confirmed |
| The aircraft-engine materials project is planned only to FY2025, and we could not find a successor | The project period on its page (FY2021 to FY2025) | Not yet confirmed |
| The new project's "confidential computing" is an extension of the problem of sharing data while keeping it secret, which the Ultra-Ultra project already tackled | The Ultra-Ultra evaluation materials say companies do not want to show their materials data to others, and report developing technology for secret yet shared use. METI's request materials also cite secure data sharing between competing firms as an aim. The request materials do not mention the Ultra-Ultra project | Inference |
| Because there are no MI-specific new calls in FY2026, NEDO's next MI projects will probably depend on the FY2027 budget | The FY2026 "materials and manufacturing" call list has no entries (checked 22 October 2026), together with the request above | Inference |
7. Glossary
- National R&D agency
- A corporation that carries out national R&D. It is not a ministry; it works under the supervision of a minister.
- Materials informatics (MI)
- Analysing data from experiments and simulations with information science to search for new materials efficiently and quickly.
- Process informatics
- Combining data science with the analysis of manufacturing processes, from small trial batches to industrial scale-up.
- Technology research association
- A body through which companies and others carry out joint R&D. In the Ultra-Ultra project, 18 companies joined the association known as ADMAT.
- Data platform
- A system that stores research data so that AI and other tools can use it. The Ultra-Ultra project built five, by purpose.
- Post-project evaluation
- The procedure in which NEDO has an outside expert committee assess a project's results after it ends.
8. References
- NEDO Organisational overview (name, headquarters, establishment, supervising minister, staff, budget and fund budgets; updated 15 April 2026) (in Japanese). https://www.nedo.go.jp/introducing/kihon.html
- NEDO Project page: Ultra-High-Throughput Design and Prototyping Technology for Ultra-Advanced Materials Development Project (FY2016 to FY2022, project code P16010, project leader, the AI tool added in FY2019) (in Japanese). https://www.nedo.go.jp/activities/ZZJP_100119.html
- NEDO Post-project evaluation of the Ultra-Ultra project: project overview, public version (20 June 2022; total cost of ¥13.8 billion as a seven-year estimate, the one-twentieth target and results, implementation structure, secret-yet-shared data technology, policy positioning) (in Japanese). https://www.nedo.go.jp/content/100947917.pdf
- NEDO Post-project evaluation report (draft) for the Ultra-Ultra project: summary (documents for the 70th meeting of NEDO's research evaluation committee, October 2022) (in Japanese). https://www.nedo.go.jp/content/100953390.pdf
- NEDO Press release: AI technology that uses multiple AIs to predict functions from complex materials data (30 June 2022) (in Japanese). https://www.nedo.go.jp/news/press/AA5_101554.html
- NEDO Project page: materials development and evaluation system for aircraft engines (FY2021 to FY2025, FY2025 budget of ¥700 million, alloy search with MI) (in Japanese). https://www.nedo.go.jp/activities/ZZJP_100196.html
- NEDO NEDO Energy and Global Environment Results Conference 2026, presentation a-05: R&D on fuel cell and water electrolysis materials using DX and materials informatics (22 July 2026) (in Japanese). https://www.nedo.go.jp/content/800068477.pdf
- NEDO NEDO Hydrogen and Fuel Cell Results Conference 2023, presentation A1-10: R&D on fuel cell materials using materials informatics (13 July 2023; period August 2020 to March 2025, planned) (in Japanese). https://www.nedo.go.jp/hydrogen2023/pdf/A1-10.pdf
- NEDO Decision on the performer of the survey on materials informatics with extensibility to process informatics, for water electrolysis hydrogen production (14 September 2022) (in Japanese). https://www.nedo.go.jp/koubo/SE3_100001_00029.html
- NEDO FY2026 calls for proposals: materials and manufacturing (the page states that no calls fall in this field; checked 22 October 2026) (in Japanese). https://www.nedo.go.jp/koubo/2026_list_05_02.html
- METI FY2027 budget request: PR materials, general account. Project on high-performance composite new materials using AI and confidential computing, ¥7.0 billion (new), with the scheme from the state to NEDO to institutes, universities and companies (in Japanese). https://www.meti.go.jp/main/yosangaisan/fy2027/pr/pdf/pr_ippan.pdf
9. Claim-to-source audit
| Claim in the article | Category | Source |
|---|---|---|
| Full name, headquarters, establishment, supervising minister, 1,640 staff, ¥505.2 billion budget, ¥11.37 trillion in fund budgets and their breakdown | Sourced | Reference 1 https://www.nedo.go.jp/introducing/kihon.html |
| The Ultra-Ultra project's period (FY2016 to FY2022), project leader and the AI tool added in FY2019 | Sourced | Reference 2 https://www.nedo.go.jp/activities/ZZJP_100119.html |
| Total cost of ¥13.8 billion (seven-year estimate), the one-twentieth target, the reported results on model materials, the structure (18 companies, 10 universities and other bodies), the FY2019 call's selections, secret-yet-shared data technology and policy positioning | Sourced | Reference 3 https://www.nedo.go.jp/content/100947917.pdf |
| The outside experts' overall assessment (largely achieved, limited range of application, data platforms not comprehensive) | Sourced | Reference 4 https://www.nedo.go.jp/content/100953390.pdf |
| The multiple-AI property prediction technique (predictions in one twenty-thousandth of the time or less) and the bodies involved | Sourced | Reference 5 https://www.nedo.go.jp/news/press/AA5_101554.html |
| Aircraft-engine materials project: period, FY2025 budget of ¥700 million, alloy search with MI | Sourced | Reference 6 https://www.nedo.go.jp/activities/ZZJP_100196.html |
| The 2026 report on fuel cell and electrolysis MI (28,850 files, performers) | Sourced | Reference 7 https://www.nedo.go.jp/content/800068477.pdf |
| The period of NIMS's part (August 2020 to March 2025, planned) | Sourced | Reference 8 https://www.nedo.go.jp/hydrogen2023/pdf/A1-10.pdf |
| The electrolysis MI survey (FY2022, one proposal, Hokkaido University) | Sourced | Reference 9 https://www.nedo.go.jp/koubo/SE3_100001_00029.html |
| No FY2026 calls in the materials and manufacturing field | Sourced | Reference 10 https://www.nedo.go.jp/koubo/2026_list_05_02.html |
| METI's FY2027 request (¥7.0 billion, scheme through NEDO) | Sourced | Reference 11 https://www.meti.go.jp/main/yosangaisan/fy2027/pr/pdf/pr_ippan.pdf |
| The four types of institution, and the arrangement of performers in Fig. 2 | Our calculation | Our own classification |
| The total spent on MI alone, approval of the FY2027 request, a successor to the aircraft-engine project, and whether the prediction technique carries over to other materials | Not yet confirmed | Not stated in references 1 to 11, or pre-approval; this article does not estimate them |
| The two inferences in the outlook (section 6): confidential computing as an extension of the Ultra-Ultra problem, and dependence on the FY2027 budget | Inference | This article's reading of references 3, 10 and 11 |
Last updated 22 October 2026 / Troy Technical
Figures come from NEDO's website, evaluation materials, press releases and conference materials, and from METI's budget request documents. Items marked "Inference" are this article's reading of the primary information, not statements by NEDO.