Hitachi, Ltd. has announced a collaboration with NVIDIA to develop multi-agent orchestration technology aimed at accelerating autonomous operations within manufacturing and social infrastructure environments. This initiative seeks to enhance Hitachi's suite of solutions, known as HMAX, by integrating AI into real-world applications.
The partnership will leverage Hitachi's social infrastructure intelligence platform, IWIM, and NVIDIA's Cosmos world foundation model to create a model for end-to-end operational optimization and autonomy. The goal is to integrate and orchestrate diverse facilities, equipment, and systems operating at customer sites, regardless of the vendor, in a secure manner.
To achieve this, Hitachi will utilize IWIM as a knowledge and reasoning foundation, integrating it with NVIDIA's Agent Toolkit software. This includes the NVIDIA OpenShell secure runtime, NVIDIA NemoClaw agent blueprint, and open-source NVIDIA Nemotron models. The collaboration aims to build a Multi-Agent Orchestration Platform to safely and securely manage multiple AI agents, even in mission-critical environments.
Hitachi and NVIDIA plan to connect IWIM with NVIDIA Cosmos world foundation models to ensure safe and autonomous control in the physical world. This will establish a reliable validation environment based on digital twins. The companies aim to build a global physical AI ecosystem centered on this platform, inviting participation from various sectors, including robotics, mobility, energy, and industrial equipment.
To support technology validation and digital transformation, Hitachi's Physical AI FDE Team will provide hands-on support to drive use case exploration and solution development. This team combines GlobalLogic's digital engineering capabilities with Hitachi's operational technology expertise.
The collaboration is part of Hitachi's broader strategy to drive AI transformation while ensuring data sovereignty. By utilizing past use cases and the AI Factory environment, Hitachi aims to address complex problems in the Physical AI domain.
The initiative will also involve technical validation of the Multi-Agent Orchestration Platform. This includes integrating knowledge regarding design and operation of social infrastructure with advanced world models, validating a supervisory agent for confidentiality and scalability, and demonstrating real-time monitoring and coordinated control of operational environments.
Hitachi's President and CEO, Toshiaki Tokunaga, expressed confidence in the collaboration's potential to realize integrated autonomous operations and accelerate the adoption of Physical AI in society. NVIDIA's Vice President of Robotics and Edge AI, Deepu Talla, highlighted the importance of connecting sensors, robots, and AI agents into systems that can understand, simulate, and act as one.