- Noetra Japan Multimodal Foundation Model powers next-generation physical AI.
- Future releases target reasoning, multimodal, and real-world AI capabilities.
A major artificial intelligence initiative has entered full-scale development after Japan-based Noetra Corp, together with its core member companies and investors including Sony Group Corporation, SoftBank Corp, NEC Corporation, and Honda Motor Co Ltd, officially launched research and development of a domestically developed multimodal foundation model. The project is intended to establish the technological base for AI-enabled robots and physical AI while bringing together a wide network of partners across Japan to strengthen domestic AI innovation.
Broad Industry Collaboration Supports AI Development
The initiative has attracted investments from 44 companies and organizations that share Noetra's long-term vision and business strategy. To accelerate progress, the company has established a dedicated research and development organization comprising engineers from its core members and investors alongside researchers from the National Institute of Advanced Industrial Science and Technology (AIST), Preferred Networks Inc, and several additional participating organizations. This collaborative structure is designed to combine industrial expertise and advanced research capabilities for the creation of a competitive homegrown multimodal foundation model.
Noetra AI Development Roadmap
| Timeline | Development Milestone |
|---|---|
| Fiscal 2026 | Launch phased reasoning foundation model development |
| Fiscal 2028 | Develop omni-modal model for text, images, video and audio |
| Fiscal 2030 | Advance Real-world Native AI with spatial understanding |
NVIDIA Infrastructure to Accelerate Model Training
To speed up model development, Noetra plans to collaborate with NVIDIA by building a large-scale AI computing infrastructure equipped with approximately 27,500 NVIDIA Rubin graphics processing units. Construction of the computing platform is scheduled to begin in April 2027, while operational deployment is expected in June 2028. The high-performance infrastructure is intended to provide the computational resources required for training increasingly sophisticated multimodal foundation models capable of supporting advanced AI applications.
Long-Term Vision for Physical AI
The company's roadmap outlines a phased approach to expanding AI capabilities over the coming years. Beginning in fiscal 2026, Noetra aims to establish a reasoning foundation model that forms the core of AI agents and natural language processing with essential foundational capabilities. By fiscal 2028, development will focus on an omni-modal foundation model capable of integrating and processing text, images, video, and audio. Looking further ahead to fiscal 2030, the objective is to realize "Real-world Native AI" capable of understanding physical properties, including spatial awareness, for deployment in real-world environments. The models will be released externally in stages based on research progress and practical implementation.
Frequently Asked Questions
What is Noetra developing through this AI initiative?
Noetra is developing a Japan-origin multimodal foundation model designed to support AI-enabled robots and physical AI through long-term research and collaboration with leading technology companies and research institutions. The roadmap includes reasoning AI, omni-modal AI capable of processing text, images, video and audio, and Real-world Native AI with spatial awareness. These models are planned for phased external release as development progresses and practical implementation milestones are achieved.
Click above to visit the official source.
Discussion
Join the conversation.