ELEMENTS Selected for "Research and Development on Data Ecosystem Construction, etc. (GENIAC)" Promoted by METI and NEDO

ELEMENTS Inc. (Headquarters: Chuo-ku, Tokyo; Representative Director: Keiki Hasegawa; hereinafter "ELEMENTS"), a provider of personal authentication AI solutions and domain-specific AI solutions, is pleased to announce its selection for the "Post-5G Information Communication System Foundation Strengthening Research and Development Project / Research and Development on Data Ecosystem Construction, etc. (GENIAC)" *3, conducted as part of the GENIAC*2 project promoted by the Ministry of Economy, Trade and Industry (METI) and the New Energy and Industrial Technology Development Organization (NEDO).

This project will focus on "tacit knowledge of skilled workers in manufacturing" to build a dataset*4. Based on the constructed dataset, collaboration is also envisioned with the "Development Project for Multimodal Foundation Models for AI Robots and Physical AI" promoted by METI and NEDO.

METI: Selected 9 new research and development themes for data ecosystem construction, etc. in the "GENIAC" project to strengthen generative AI development capabilities and social implementation (July 2, 2026) https://www.meti.go.jp/press/2026/07/20260702001/20260702001.html

NEDO: Decision on the implementation structure for the "Post-5G Information Communication System Foundation Strengthening Research and Development Project / Research and Development on Data Ecosystem Construction, etc. (GENIAC)" (July 2, 2026) https://www.nedo.go.jp/koubo/CD3_100430.html

*1 A virtuous cycle where data from various fields is collected, used for AI development and provision, and further data is collected through the widespread use of that AI.

*2 GENIAC (Generative AI Accelerator Challenge): A project promoted by METI and NEDO aimed at strengthening Japan's generative AI development capabilities and social implementation.

*3 Regarding the public offering for the "Post-5G Information Communication System Foundation Strengthening Research and Development Project / Research and Development on Data Ecosystem Construction, etc. (GENIAC)" (March 25, 2026) https://www.nedo.go.jp/koubo/CD2_100430.html

*4 A collection of data gathered and organized for a specific purpose. In AI machine learning and statistical analysis, it refers to a set of information directly used for model training and evaluation.

*5 AI models that can handle diverse data such as voice, images, videos, and sensor data, not limited to language.

◾️Summary

・ELEMENTS has been selected as an implementing contractor for the "Post-5G Information Communication System Foundation Strengthening Research and Development Project / Research and Development on Data Ecosystem Construction, etc. (GENIAC)" publicly offered by NEDO from March 2026.

・This project will focus on "tacit knowledge of skilled workers in manufacturing" to build a dataset.

・Recently, the "depletion of learning data" on the web is imminent, and the utilization of "proprietary data" held by companies and organizations, which is not available on the web, will become increasingly important.

・The manufacturing sector is a treasure trove of AI learning data for Japan, which has strengths in manufacturing. ELEMENTS will convert and standardize this information into a format that AI foundation models can handle, operating it as a domestic data infrastructure that contributes to economic security and does not rely on overseas companies.

・The cross-company data infrastructure to be built will be handled within ELEMENTS CLOUD, a highly secure environment operated by ELEMENTS that enables efficient learning through GPU management.

・Furthermore, this data is expected to be linked with the "Development Project for Multimodal Foundation Models for AI Robots and Physical AI" promoted by METI and NEDO, and with physical AI-related projects.

◾️Background

Until now, generative AI has learned from large amounts of text data on the internet to improve its performance. However, the "depletion of learning data" on the web is imminent, and the utilization of "proprietary data" held by companies and organizations, which is not available on the web, will become increasingly important.

In the manufacturing industry, publicly available data on the web is limited, and to utilize the "proprietary data" held by companies and organizations, it is necessary to convert and standardize the data into a format that AI foundation models can handle.

In many companies, data such as design documents, specifications, quality trouble reports, and defect history related to manufacturing are dispersed within the company in an unstructured and non-standardized state.

Leveraging its accumulated technology and knowledge, ELEMENTS will create "structured datasets" from companies' "proprietary data" that AI can read and learn from, and create "multimodal datasets" from physical information in the manufacturing process, such as human actions.

In addition, ELEMENTS has built its own foundation for large-scale learning and high-speed inference in cloud services that handle personal information, including personal authentication. This knowledge, along with operational know-how for security infrastructure that has served over 700 companies, including financial institutions where robustness is paramount, is available.

Utilizing the general-purpose know-how gained from AI data processing, learning, evaluation, and GPU utilization efficiency in these personal authentication businesses, ELEMENTS will engage in AI model evaluation and domain-specific (vertical) AI businesses.

Reference: METI will launch the "Development Project for Multimodal Foundation Models for AI Robots and Physical AI" (June 30, 2026) https://www.meti.go.jp/press/2026/06/20260630005/20260630005.html

◾️Initiative Details

In this project, focusing on "tacit knowledge of skilled workers in manufacturing," the following datasets will be constructed. The constructed datasets will be promoted for collaboration with the "Development Project for Multimodal Foundation Models for AI Robots and Physical AI" promoted by METI and NEDO.

1. Structured dataset integrating design and manufacturing data

Internal documents such as quality trouble reports, design specifications, inspection records, and drawings will be digitized through structural analysis, OCR, and semantic extraction. After anonymization and masking of proper nouns and filtering of confidential information, they will be converted and standardized into a format that AI foundation models can handle.

2. Multimodal human action dataset capturing skilled workers' operations

First-person camera (a shooting method that records scenery and actions from the same position and height as the "eyes" of the photographer) and third-person camera (a camera work or system that captures the whole or the back from a slightly distant third-person perspective) will be used to film the hand and full-body movements of skilled technicians, along with explanatory audio capturing key decision-making points. By adding meaning to tacit knowledge through audio commentary explaining "why that action is taken," a unique data asset will be built that connects actions, visuals, and intentions.

◾️Future Development

In this project, ELEMENTS will collaborate with data holders from multiple major manufacturing companies to formalize the "tacit knowledge" accumulated in Japanese manufacturing sites into multimodal datasets that AI can learn from.

These datasets will serve as the core of a "data ecosystem" connecting data providers, AI developers, and manufacturing sites, and will be provided fairly to domestic businesses that meet the requirements. They will be operated as a domestic data infrastructure that contributes to economic security and does not rely on overseas companies, with collaboration envisioned with physical AI-related projects promoted by METI and NEDO.

Furthermore, this project has been adopted as a GX framework aiming for the simultaneous realization of CO2 emission reduction and economic growth.

ELEMENTS aims to contribute to the improvement of productivity and the reduction of environmental impact in the entire manufacturing industry through the reduction of defective products and the improvement of energy saving performance by optimizing design.

◾️About ELEMENTS CLOUD

This is a service that constructs cloud environments and operates data centers using a suite of software developed over more than 10 years of AI service development and provision by the ELEMENTS Group, enabling secure, large-scale learning and high-speed inference environments.

Web: https://lp.elements-cloud.jp/

◾️About ELEMENTS Inc.

With the group mission "BEYOND SCIENCE FICTION," ELEMENTS is a company that develops three solutions: personal authentication, personal information management, and personal optimization. It aims to solve social issues caused by financial crime, mass production, and mass disposal. Its current flagship service, the online identity verification service, has been introduced to over 700 companies in a wide range of industries, including finance and telecommunications.

Location: 5F, Nihonbashi Life Science Building 3, 3-8-3 Nihonbashi-Honcho, Chuo-ku, Tokyo

*The headquarters will be relocated to the following address on July 6, 2026.

5F, Nihonbashi Life Science Building 6, 3-9-4 Nihonbashi-Honcho, Chuo-ku, Tokyo

Representative: Keiki Hasegawa, President and Representative Director

Stock Code: Tokyo Stock Exchange Growth Market 5246

Established: December 2013

Business Description: Development and provision of personal authentication solutions utilizing biometric authentication, image analysis, and machine learning technologies; personal optimization solutions in food, clothing, and housing; and cloud services for managing personal information.

Website: https://elementsinc.jp/

*Company names and product/service names mentioned in this press release are registered trademarks or trademarks of their respective companies.