TOKYO – June 16, 2026 – Preferred Networks, Inc. (PFN) today announced its investment in Noetra Corp. (Noetra) and its participation in Noetra’s project focused on Japan-developed multimodal foundation model for Physical AI.

This initiative follows the selection of Noetra and the National Institute of Advanced Industrial Science and Technology (AIST) for "Project for Developing Multimodal Foundation Model Toward AI Robots and Physical AI," a national project sponsored by Japan’s Ministry of Economy, Trade and Industry (METI) and the New Energy and Industrial Technology Development Organization (NEDO). Within this project, Noetra will develop a Japan-made multimodal foundation model and related technologies that will serve as the core platform for AI robotics and Physical AI.

Daisuke Okanohara, co-founder and CEO of PFN, will lead model development at Noetra as the Joint R&D Project Director for the NEDO initiative. He will also serve as a Technical Director for the exploratory research track spearheaded by AIST. Engineers from PFN’s PLaMo™ development team, including Daisuke Tanaka, who leads the development of PLaMo and PFN’s other generative AI foundation models, will be seconded to Noetra to drive its model development forward.

PFN has a proven track record of developing its own generative AI foundation model PLaMo fully from scratch. Additionally, the company has commercialized advanced robotics technologies, such as the autonomous mobile robot Kachaka, and real-world control systems for complex environments including oil refineries. By leveraging its comprehensive AI technology stack spanning generative AI, computing infrastructure, and AI chips alongside its expertise in robotics, automation, and optimization, PFN will actively contribute to the development and practical application of Noetra’s multimodal foundation model.

Physical AI is expected to be adopted across a wide range of industries including manufacturing, logistics, infrastructure, mobility, and robotics by integrating and understanding diverse real-world data types such as images, videos, audio, sensor data, and physical properties to execute recognition, reasoning, planning, and control. Building a secure domestic multimodal foundation model that leverages Japan's unique strengths – such as frontline industrial data, a robust manufacturing foundation, and deep industry expertise – is a critical step toward strengthening the country’s industrial competitiveness.

Moving forward, PFN will collaborate closely with participating companies and related institutions to advance Noetra's Japan-developed multimodal foundation model project, which will underpin the real-world deployment of Physical AI.

Daisuke Okanohara, Co-founder and CEO of PFN said:

"We see tremendous value in contributing to the Japan-made multimodal foundation model project by Noetra and AIST. PFN will contribute our expertise in solving real-world challenges through AI, our experience in developing the PLaMo generative AI foundation model from scratch, and our knowledge of robotics. In close collaboration with the participating companies at Noetra, we are fully committed to advancing Physical AI by building upon Japan's core strengths in frontline operational capabilities and its world-class manufacturing foundation."