NVIDIA CEO Jensen Huang Visits Japan, Signs Three Major Agreements to Transform Japanese Industries

NVIDIA’s CEO Jensen Huang visited Tokyo on July 15 and 16, meeting with leaders from Japan’s industrial sector and semiconductor supply chain. This visit, following a keynote speech in Taiwan a few weeks earlier and a trip to South Korea months ago, resulted in several major agreements spanning across Japan’s tech ecosystem, according to a report by Kate Park of TechCrunch AI.

Thirty years ago, video game company Sega supported NVIDIA with a $5 million investment when the company was on the brink of bankruptcy. Today, NVIDIA has grown into one of the world’s leading semiconductor companies by market capitalization and is once again forging a symbiotic relationship with Japanese industrial giants. At the heart of this transformation is the construction of the “physical AI era,” with three projects taking the lead.

Breakdown of the Three Major Agreements

Huang’s visit to Japan resulted in three key agreements: first, the construction of a national-scale AI factory; second, collaboration with major domestic robotics companies; and third, agreements with suppliers of semiconductor materials for next-generation AI chips. NVIDIA is targeting Japan’s factory floors, and many of the country’s leading manufacturers have emerged as potential collaborators.

In a press conference, Huang emphasized that applying AI to factory production lines and logistics warehouses represents the next frontier for growth. This marks a shift for NVIDIA from its traditional business model, which has relied heavily on selling GPUs for data centers, toward a full-fledged entry into the industrial automation market. Given Japan’s global leadership in machine tools, industrial robots, and automobile production, the country appears to be an ideal field for implementing physical AI.

Noetra: Japan’s Sovereign AI Strategy

The Japanese government is reluctant to rely on American or Chinese AI to operate its factories and robots. In response, it has brought together around 44 domestic companies, with SoftBank, Sony, NEC, and Honda as core members, to form a consortium called Noetra. This organization aims to develop proprietary AI technologies for robots, vehicles, and factory floors.

Tokyo plans to invest up to ¥1 trillion (approximately $6.2 billion) over the next five years. These funds will be directed toward the development of domestic “physical AI,” a foundational model designed to operate machinery. While Japan seeks to maintain sovereignty over AI software, NVIDIA will remain a key player in building the necessary hardware.

NVIDIA plans to construct the Vera Rubin AI Factory, a massive data center expected to begin operations in 2028 and equipped with NVIDIA’s next-generation chips. The facility will house 13,750 Vera CPUs and 27,500 Rubin GPUs, consuming 140 megawatts of electricity. Noetra will oversee the entire project, including the data center’s construction plan.

Noetra’s roadmap includes three phases. The first phase aims to develop a reasoning model with advanced Japanese language capabilities by fiscal 2026. The second phase, to be completed by 2028, will focus on creating an “omnimodal” version capable of processing text, images, video, and audio. Finally, the third phase, targeting completion by 2030, seeks to develop “Real-world Native AI” capable of controlling robots, which will then be gradually opened to external Noetra developers.

A notable feature of this three-phase approach is its focus on industrial applications rather than general-purpose AI. Starting from a reasoning model specialized in Japanese language processing, it progresses to multimodal capabilities and ultimately culminates in robot control. The phased release strategy appears to be designed to allow domestic suppliers and system integrators to adapt gradually.

Robotics Alliance: The Move to Cosmos

NVIDIA has set its sights on Japan’s factory floors, and many of the country’s leading robotics and manufacturing firms have responded. Companies such as Fanuc, Yaskawa Electric, Kawasaki Heavy Industries, Fujitsu, Hitachi, NEC, Sony, SoftBank, Kubota, and the robotics organization AIRoA have announced their participation in development efforts based on NVIDIA’s Cosmos model.

Cosmos, an open-model project launched by NVIDIA in collaboration with select global AI research institutions in May 2025, is designed as a physical-world understanding model for autonomous driving and robot control. NVIDIA aims to use this model as a cornerstone for building an industrial AI ecosystem.

During his Tokyo visit, NVIDIA unveiled Cosmos 3 Edge, a new version of the model designed to operate within robots themselves using the Jetson Thor chip. By enabling inference directly on edge devices, this model reduces dependency on cloud computing and meets the low-latency demands of factory floors.

Some companies have already begun testing shared control systems. Honda R&D and Omron are currently advancing developments using these tools. Huang remarked, “The next frontier of AI lies in the physical world. This is a once-in-a-century opportunity for Japan. Japan invented modern manufacturing. Now it has the chance to reinvent it for the era of intelligent industries.”

For Japanese robotics manufacturers, transitioning to Cosmos 3 Edge would mean a shift from their existing control architectures. Companies like Fanuc and Yaskawa Electric, which have long developed proprietary control systems, now face the challenge of standardizing on NVIDIA’s platform—a shift that will likely shape the competitive landscape for Japan’s robotics sector.

Transforming the Semiconductor Supply Chain

The third pillar of the agreements involves strengthening partnerships with suppliers of materials for next-generation AI chips. Japan holds a dominant share of the global market for semiconductor materials and manufacturing equipment, particularly in areas like chemical mechanical polishing (CMP) materials, photoresists, and high-purity gases.

Producing NVIDIA’s next-generation Rubin GPUs and Vera CPUs requires cutting-edge packaging technologies and materials. By establishing direct contracts with Japanese material suppliers, in addition to its existing Taiwanese partner TSMC, NVIDIA aims to mitigate geopolitical risks and stabilize its supply chain.

This development presents an opportunity for Japan’s semiconductor materials industry. As the miniaturization of semiconductors approaches its limits, innovations in materials and manufacturing equipment are becoming increasingly critical to improving chip performance. Collaborating directly with NVIDIA offers Japanese suppliers a chance to set the trajectory for future technological advancements.

Industrial Transformation Through Physical AI

Huang’s vision for physical AI goes beyond merely enhancing robot control. It involves integrating AI into every stage of factory operations—from design and production planning to quality control and maintenance forecasting—optimizing these processes holistically. NVIDIA’s Cosmos model is expected to serve as the common foundation for achieving this integration.

In the automotive sector, Honda has already begun development based on the Cosmos model, with applications ranging from autonomous driving control to production line optimization. Kubota, a manufacturer of machine tools, is also participating, considering applications in agricultural and construction machinery.

The Japanese government’s decision to invest ¥1 trillion in Noetra underscores the importance of ensuring sovereignty in industrial AI implementation. Just as the European Union established data sovereignty through GDPR, Japan is accelerating the creation of frameworks to secure its industrial AI sovereignty.

Editorial Opinion

Jensen Huang’s visit to Japan signifies a new phase in the relationship between NVIDIA and Japan, moving beyond semiconductor procurement into the realm of industrial restructuring. In the short term, the development of Noetra’s Japanese language reasoning model is expected to intensify the competition for domestic AI talent. With a shortfall of hundreds of thousands of AI engineers annually, this could lead to rising labor costs and delays in project timelines. Additionally, the rollout of Cosmos 3 Edge may compel small and medium-sized factory automation equipment manufacturers to transition away from existing control systems. If the integration of a standardized platform progresses, it could create a new competitive landscape for Japanese robotics manufacturers who have traditionally relied on proprietary technologies.

In the long term, if Japan succeeds in establishing a foundational physical AI model, it could position itself as a third pole distinct from the U.S. and China. However, the decision-making speed of the 44-member Noetra consortium remains a classic challenge, as past Japanese consortium projects have rarely succeeded in securing technological dominance. Furthermore, the Vera Rubin AI Factory’s projected electricity demand of 140 megawatts could have a direct impact on Japan’s energy policies.

References

• “What to watch for after Jensen Huang’s Japan visit”, by Kate Park — TechCrunch AI, 2026-07-19T21:16:07.000Z (ARR)
• Source URL: https://techcrunch.com/2026/07/19/what-to-watch-for-after-jensen-huangs-japan-visit/