
Report: Japan Spends $2.4 Billion on NVIDIA Rubin Chips to Build Robotic AI Systems
Japan is making a bold $2.4 billion investment in NVIDIA's next-generation Rubin chips, establishing the state-led Noetra to integrate the AI capabilities of dozens of companies, including SoftBank, Toyota Motor Corp.'s Preferred Networks, and NEC. The goal is to capture 30% of the 60 trillion yen global robotics market by 2040
Japan is betting on next-generation artificial intelligence chips to build an autonomous AI foundation for its domestic robotics industry.
According to Bloomberg, Japan plans to purchase 27,500 next-generation Rubin chips from NVIDIA, with a total investment of 387.3 billion yen (approximately $2.4 billion), to build large-scale data centers and develop local foundational AI models for robotics. The newly established company responsible for coordinating this project, Noetra Corp., has received the aforementioned government funding, which must be used by March of next year. NVIDIA CEO Jensen Huang stated on Wednesday that Japan has many excellent ideas but suffers from a severe labor shortage, adding that "the economy can take off again through automation, AI, and robotics technology."
Dozens of companies, including SoftBank Group Corp., Preferred Networks Inc. (a subsidiary of Toyota Motor Corp.), and NEC Corp., have participated in the establishment and operation of Noetra. Noetra plans to release its first AI model by March of next year and will subsequently launch specialized models tailored for robotic application scenarios over the following years.
State-Led Initiative to Consolidate Fragmented AI Efforts
Noetra was established under the leadership of the Japanese government, with a core mission to integrate AI research and development efforts previously scattered across various companies. Hironobu Tamba, the president of Noetra, who previously led the development of SoftBank's large language models, stated, "Our goal is to create a true third option—a solution that Japan and other countries can choose."
The companies participating in this project have each accumulated significant AI R&D foundations. SoftBank possesses the Sarashina large language model, Preferred Networks has developed the PLaMo model, NEC's flagship model is named cotomi, and Fujitsu also has its own independent AI model strategy. Noetra will draw engineers from these companies to collaboratively advance the project's implementation.
Although the scale of this procurement is considerable, it remains relatively limited compared to Microsoft's plan to eventually build data centers equipped with hundreds of thousands of NVIDIA Vera Rubin chips.
Robotics Ambition: Capturing 30% of the Global Market by 2040
Developing domestic physical AI models is a core pillar of Japan's broader strategy to establish itself as a global leading center for the AI and robotics industries. The Japanese government has set a target to capture more than 30% of the global robotics market, estimated to be worth 60 trillion yen, by 2040.
As one of the world's most important manufacturers of industrial robots, Japan's industrial endowment provides a realistic foundation for developing robotic AI. At the same time, the continuous population decline and severe labor shortage make the promotion of AI and automation not an option, but a necessity driven by reality. Noetra's long-term goal is to launch customized AI models specifically serving robotic application scenarios within a few years.
Reducing Foreign Dependence and Strengthening Technological Autonomy
This initiative to build an autonomous AI system is part of a series of actions by the Japanese government to reduce dependence on foreign technology and strengthen national security. Facing an increasingly competitive landscape in global AI model development, Japan intends to integrate industrial strength to construct a third pole distinct from the US and Chinese AI ecosystems.
Tamba stated that the original intention behind establishing Noetra was to break the long-standing fragmentation in Japan's AI sector, consolidating dispersed efforts into a competitive, cohesive force. The company will complete the release of its first AI model by March of next year and continue to iterate and update it.
