After Huang In-hoon ' s two-day trip to Tokyo, the map of Ying Weida ' s cooperation in Japan was further clarified. The focus of this landing is not just the sale of chips, but also national AI infrastructure, robotic model deployment, and deeper access to automobiles and factory systems. For Japan, this is also part of its plan for manufacturing upgrading and AI autonomy.

Japan launched the national AI plant

Japan is promoting a local AI project called Noetra, with core participants including soft silver, Sony, NEC and Honda, with the objective of developing national models for robotics, automobiles and plant scenes. According to reports, Japan plans to invest up to 1 trillion yen over five years in “physical AI”, which can be directly driven.

Within this framework, Weida will build an AI plant based on the Vera Rubin architecture, a large data centre. The project is expected to start in 2028, with 13,750 Vera CPUs and 27,500 Rubin GPUs of approximately 140 MW power. Noetra will be responsible for the overall advancement and plans to build the data centre.

The route of Noetra is divided into three phases: first, the introduction of a model of reasoning for Japanese-speaking abilities in fiscal year 2026; then, around 2028, the expansion of multimodular versions of text, image, video and audio that can be processed; and, by 2030, the advancement of “real-world primary AI” for directly operating robotic systems, which is gradually open to outside developers.

The robot business gathers around Cosmos

The second focus of the tour was to bring Japanese manufacturing and robotics into their Cosmos model ecology. It was reported that Hanaco, Angawa Electrician, Kawasaki Heavy Industries, Fujitung, Higashi, NEC, Sony, Soft Silver, Kubuda and the robotic organization AIROA, all planned to be developed on the basis of the Cosmos model.

Weeda also released Cosmos 3 Edge in Tokyo. This version can be operated on the Jetson Thor chip and deployed directly within the machine and equipment, rather than relying solely on clouds. For plant robotics and industrial equipment, this means that the model can operate closer to the terminal.

Some enterprises are already testing shared control systems, and the Honda R & D and Omlon companies have begun to use tools. It is the hope of YVD to further the application of the generated AI from the data centre to the factory workshop and robotics.

Cooperation with Toyota extended to plant and vehicle software

Toyota is another key partner in Huang In-hoon's trip. Toyota had previously used the British Weeda chip in a garage of automobile technology and committed in January 2025 to the next generation of vehicles using the Nvidia Drive platform. The latest cooperation further extends to manufacturing and transport systems.

Reports indicate that the scope of cooperation between the parties has covered production line simulations, vehicle operating software and road traffic identification systems. Toyota remains cautious on the automatic driving path, with the current focus on advanced assistive driving systems requiring the participation of drivers rather than a fully artificially weakened programme.

Japan integrates physics AI into industry strategy

Hwang In-hoon's visit revealed a greater signal that Japan is placing physics AI at the centre of industrial policy. In the face of a contraction of the labour force, Japan hopes to deploy 10 million robots with AI in 18 industries by 2040, with approximately $65 billion in public and private investment.

Japan's AI Robotics Strategy, released in March this year, proposes that by 2040 Japan will have managed to take over 30 percent of the global AI Robotics market. The Government of Japan estimates that this market is about 20 trillion yen. The Ministry of Economy, Trade and Industry is also supporting the development of indigenous basic models, with a view to maximizing the availability of data, algorithms and industrial scenarios within the national system.

However, while promoting AI autonomy in Japan, bottom-level computing remains highly dependent on Weida. In other words, Japan wanted to establish a more independent AI industrial system, but at this stage the United States chip was still needed.