J-POWER, Hitachi, and Macnica have agreed to jointly study the construction of a physical AI development platform with the aim of implementing physical AI in critical social infrastructure sectors.
Figure 1. Loop structure image of the AI factory (※Figure created by generative AI)
J-POWER (Electric Power Development Co., Ltd.), Hitachi, Ltd. (Hitachi), and Macnica (Macnica) have agreed to jointly study the construction of a physical AI development platform with the aim of implementing physical AI in critical social infrastructure fields.
In recent years, in critical social infrastructure sectors such as electricity, expectations are rising for physical AI to handle the sophistication and automation of inspection and maintenance tasks, due to a shortage of skilled workers and the aging of equipment.
This time, with a view to implementation on the AI-DC (AI-DC) *1 that J-POWER and Hitachi are jointly considering, we will explore the construction of a physical AI development platform utilizing NVIDIA's advanced AI technology, based on the concept of AI Factory *2. On this platform, we will enable a development cycle (Sim2Real *3) in which AI is trained in a simulation environment to learn infrastructure operation, maintenance, and control, and the verified AI model is installed on equipment and robots for execution on-site, while the obtained data is used for further learning. This will aim to realize autonomous decision-making and advanced work support on-site. As the first initiative, we will start exploring demonstrations targeting next-generation power transmission line inspection drones, utilizing real-world fields such as power generation and transmission facilities owned by the J-POWER Group.
In the future, through the physical AI development platform built in this initiative, we aim to establish a development and operation process that connects digital twins, simulations, AI, and actual equipment, and to realize an environment on the planned AI-DC where AI learning, verification, and operation can be carried out in an integrated manner. Through these efforts, we will promote the deployment of this technology to use cases in various critical social infrastructure fields, including operation, inspection, and maintenance.
*1 In the press release dated July 7, 2025, "J-POWER and Hitachi Agree to Jointly Study the Construction of a Data Center for AI for J-POWER," they are considering energy optimization through the realization of a distributed AI-DC and the construction of a highly reliable and secure environment.
*2 AI Factory: A concept for an AI development and operation platform that supports data collection, learning, validation, inference, and continuous improvement.
*3 Sim2Real (Simulation to Reality): An approach that applies an AI model trained and validated in a simulation environment to a real environment, and then uses the real-world data to train and validate it again.
Figure 2. Image of the demonstration of a next-generation power line inspection drone (※Figure created using AI generation)
background
With advancements in generative AI, AI technology is evolving beyond digital applications to include physical AI that operates in the physical world, such as robots and drones. In critical social infrastructure sectors, its use is expected to enhance the sophistication and efficiency of equipment inspection and maintenance operations, as well as improve safety.
On the other hand, when it comes to the practical application of physical AI, there are constraints in terms of time, cost, and security if AI models are trained and validated solely in real-world environments. Therefore, it is necessary to train and validate models on a digital twin that replicates the real world, apply the results to actual devices, and then use data obtained in real-world environments to further refine and improve the training process. Furthermore, in order to continuously develop and operate physical AI, a platform that can safely manage highly confidential data while conducting AI training, validation, and operation is essential.
Details of the initiative and the division of roles among the three companies
J-POWER, Hitachi, and Macnica will combine their respective expertise, technologies, and assets to jointly explore the construction of a physical AI development platform. This development platform will build an environment for learning and verifying physical AI, with a view to utilizing advanced AI and simulation technologies such as the NVIDIA Omniverse Library *4, NVIDIA Cosmos *5, and NVIDIA Isaac *6. In addition, leveraging the knowledge gained from Hitachi and NVIDIA's collaboration *7 to build a physical AI ecosystem, they will combine Hitachi's "HMAX Data Fabric" *8, an AI data linkage and utilization platform, and "HMAX AI Operations" *9, which supports the safe and stable operation of AI models, with the aim of establishing a development cycle that connects digital twins, simulations, AI, and actual machines.
Furthermore, we will develop an edge AI execution environment utilizing the NVIDIA Jetson *10 edge AI computing platform, and proceed with the study of a robot intelligence platform that can perform inference processing such as equipment recognition, anomaly detection, autonomous movement, and path planning in real time on-site. This will enable the realization of physical AI that operates safely and reliably at critical social infrastructure sites without relying on constant connectivity to the cloud.
|
company |
Main Role |
|
J-POWER |
Planning use cases, providing real-world field experience, and evaluating and verifying the results. |
|
Hitachi |
Our FDE (Forward Deployed Engineers) team, possessing domain knowledge in critical social infrastructure such as power, as well as expertise in IT, OT, and AI, will concretize and evaluate use cases for physical AI, consider the architecture of the development platform, and build and verify it. |
|
Macnica |
Design, construction, and operation support for Sim2Real proof-of-concept environments; support for exploring use cases and applications such as robotics. |
*4 NVIDIA Omniverse: A collection of modular physical AI simulation libraries that developers and AI agents can integrate into existing applications and workflows.
*5 NVIDIA Cosmos: A platform consisting of a generative world foundation model, guardrails, and accelerated data processing and curation pipelines to drive the development of physical AI systems.
*6 NVIDIA Isaac: An open robotics development platform that integrates simulation and robot learning frameworks, libraries, AI models, and reference workflows, used to build autonomous machines, robots, and humanoids.
*7 Press release dated July 16, 2026: "Hitachi and NVIDIA collaborate to achieve overall site autonomy and advance HMAX through integrated control of physical AI"
*8 HMAX Data Fabric: A data platform that integrates and systematizes diverse on-site data and the tacit knowledge of experts, and stores and shares it as knowledge assets that can be understood and utilized by AI.
*9 HMAX AI Operations: A solution that continuously monitors and mitigates AI operational status and specific risks, supporting safe and reliable operations.
*10 NVIDIA Jetson: An edge AI computing platform for performing AI inference and autonomous control on field equipment such as robots, drones, and autonomous machines.
Future developments
The three companies will begin with the verification results of their first power line inspection drone project, and will proceed with practical applications aimed at enhancing on-site safety, from automating inspection and maintenance work to advanced autonomous control and disaster recovery support. Furthermore, through the future implementation of a physical AI development platform on AI-DC, they aim for the integrated operation of the digital infrastructure supporting AI and the power infrastructure, leading to a further deepening of the "Watt-Bit Collaboration" *11, 12, 13. In addition, they will promote partnerships and open innovation not only in the power sector but also in various critical social infrastructure fields such as railways and various industrial plants, contributing to addressing the worsening labor shortage and maintaining and developing sustainable and resilient social infrastructure.
*11 Watt-bit collaboration: Effective collaboration of power and communication through efficient power and communication infrastructure (a coined term combining the power unit "W (watt)" and the information and communication unit "bit").
*12 Press release dated April 22, 2026: "Agreement to jointly study workload shifting and wide-area optical network technologies with the aim of integrated operation of distributed AI data centers."
*13 Press release dated October 8, 2026: "We have commenced the Ministry of Internal Affairs and Communications' FY2026 'Watt-Bit Linkage Related Demonstration Project'."
Comment from Shingo Koizumi, Executive Officer of J-POWER
The J-POWER Group owns and operates a diverse range of power generation and transmission/distribution facilities both domestically and internationally, supporting the stable supply of electricity for many years. In critical social infrastructure in Japan, where labor shortages and aging facilities are becoming challenges, automation using AI and robotics, as well as improvements in safety and productivity, are becoming increasingly important. This time, we will combine the technological capabilities of Hitachi, Ltd. and Macnica, along with the knowledge, data, and diverse demonstration fields that the J-POWER Group has cultivated through more than 70 years of operating power facilities, and actively incorporate NVIDIA's advanced technologies and expertise to accelerate the social implementation of physical AI by conducting verification with a view to application to actual facilities and operations. Moving forward, we will accumulate the knowledge and results obtained through this initiative and develop it into a "lighthouse" (a landmark model) that leads the use of physical AI in critical social infrastructure, and contribute to the realization of sustainable infrastructure operation.
Comment from Yuichi Yagawa, General Manager of the Social Systems Business Division, Hitachi, Ltd.
Hitachi has been working with J-POWER to build AI-DCs and realize watt-bit integration. This joint study is seen as an important step in developing these efforts and leading to the creation of a concrete use case for physical AI and its social implementation. Leveraging its expertise in OT, IT, and AI, Hitachi will contribute to improving safety and productivity in critical social infrastructure sectors by building a physical AI development platform that supports everything from AI development and verification to on-site application. Furthermore, leveraging its collaboration with NVIDIA to build a physical AI ecosystem, Hitachi will promote further advancements in AI utilization and value creation through HMAX.
Comment from Keisuke Ito Macnica CLAVIS Company
As an authorized distributor of NVIDIA, Macnica has been working to incorporate cutting-edge technologies and solve challenges in the fields of AI and robotics. We see this joint study as an important step in creating concrete use cases for physical AI and leading to its social implementation through the Sim2Real demonstration environment, which connects digital twin simulation with actual machines. By supporting the design, construction, and operation of the demonstration environment and the consideration of use cases and applications, Macnica will combine the knowledge, data, and demonstration fields of the J-POWER Group with Hitachi's expertise in OT, IT, and AI, and Macnica 's technological capabilities to contribute to improving safety and productivity in critical social infrastructure and realizing sustainable infrastructure operation.
Comment from Marc Spieler, Lead for Global Energy at NVIDIA
Physical AI will transform the way critical energy infrastructure is inspected and maintained, contributing to addressing labor shortages and improving safety and reliability. Hitachi, J-POWER, Macnica are combining their respective industry expertise with NVIDIA's full-stack AI platform to advance the learning and validation of intelligent machines in simulation environments and promote their deployment in actual operations.
About J-POWER
J-POWER is an electric power company founded in 1952. The J-POWER Group's mission is to continuously provide the energy that people need and contribute to the sustainable development of Japan and the world. To this end, it has been engaged in power generation and transmission/distribution businesses using hydroelectric, thermal, wind, geothermal, and solar power both domestically and internationally. In order to achieve this mission, the J-POWER Group has announced "J-POWER BLUE MISSION 2050" and is striving to achieve carbon neutrality in its power generation business by 2050. It is promoting initiatives such as accelerating the development of CO2-free power sources, upcycling existing thermal power plants, and generating, manufacturing, and supplying CO2-free hydrogen, with the goal of reducing emissions by 22.5 million tons by 2030* and achieving net-zero emissions by 2050.
*All figures are compared to fiscal year 2013, and represent J-POWER's domestic power generation business.
About Hitachi, Ltd.
Hitachi aims to be a global leader in continuously innovating social infrastructure digitally through its Social Innovation Bonds (SIB) business, leveraging IT, OT (Operational Technology), and products, contributing to the realization of a Harmonized Society where environment, well-being, and economic growth are in harmony. In addition to its four sectors—Digital Systems & Services, Energy, Mobility, and Connected Industries—Hitachi operates globally through a strategic SIB business unit that creates new growth businesses. With Lumada at its core, Hitachi creates value from data to solve the challenges of its customers and society. For fiscal year 2025 (ending March 2026), Hitachi projected net sales of 10,586.7 billion yen. As of the end of March 2026, Hitachi has 606 consolidated subsidiaries and approximately 290,000 employees worldwide. For more information, please visit www.hitachi.com/ja-jp/.
About Macnica
Macnica is Service & Solution Company that handles the latest technologies comprehensively, with semiconductors and cybersecurity at its core. Operating in 100 locations across 33 countries/regions worldwide, we leverage our technological expertise and global network cultivated over more than 50 years of history to discover, propose, and implement cutting-edge technologies such as AI, IoT, and autonomous driving. For more information, https://www.macnica.co.jp/ Please take a look.
Inquiries from the press regarding this matter
Macnica://www.macnica.co.jp
Public Relations Office Miyahara E-mail: macpr@macnica.co.jp
Macnica Building 1, 1-6-3 Shin-Yokohama, Kohoku-ku, Yokohama 222-8561