NVIDIA founder and CEO Jensen Huang visited Tokyo to meet with Japanese government officials, technology leaders, and supply chain partners as the country expands its physical AI, robotics, and computing infrastructure, according to reporting by Hugging Face.
Japan launched a government-backed Physical AI Initiative led by Economy Minister Ryosei Akazawa and Huang at Prince Park Tower. The initiative will develop open multimodal foundation models for AI agents, digital twins, and robotics applications. Huang also met with the CEOs of Fujitsu, Kawasaki Heavy Industries, FANUC, and Yaskawa to integrate physical AI directly into factory floors. Separately, Huang hosted more than 30 executives from 16 supply chain firms—including Advantest, Tokyo Electron, Kyocera, Mitsubishi Electric, Panasonic, and Renesas—in Tokyo's Kanda district.
Network and industrial updates
SoftBank Corp. aligned with NVIDIA on its next phase of AI-native network collaboration. SoftBank is deploying GB200-class AI infrastructure, NVIDIA AI Aerial platforms, and Nemotron models alongside SB Intuitions' Sarashina generative AI model. Meanwhile, industrial partners such as OMRON, Fujitsu, Hitachi, Shimizu Corporation, and Asilla are adopting updated NVIDIA Metropolis tools—including VSS Blueprint 3.2 and DeepStream 9.1—to build vision AI agents for factories, construction sites, and public spaces.
Healthcare and science
Japanese healthcare providers are incorporating accelerated computing directly into medical systems. Canon launched Japan's first NVIDIA-accelerated photon-counting CT system, and Fujifilm commercialized a whole-body CT system powered by NVIDIA Blackwell chips. In biopharma, drug discovery consortium Tokyo-1 added Eisai alongside existing members Astellas, Daiichi Sankyo, and Ono Pharmaceuticals. SyntheticGestalt introduced two molecular models, ZAO and KOYA, that interface with the NVIDIA BioNeMo Agent Toolkit.
At research institute RIKEN, two NVIDIA-powered supercomputers entered operation. RIKYU deploys 1,600 Blackwell GPUs on the GB200 NVL4 platform for AI foundation models across life and materials sciences. ROQUO integrates 540 Blackwell GPUs with Quantinuum's trapped-ion Reimei quantum computer for hybrid quantum-HPC workloads using CUDA-Q.
Financial infrastructure
Japanese financial institutions are deploying dedicated AI factories on premises. Mizuho plans to build an AI factory starting with NVIDIA DGX B200 systems to run autonomous workflows securely. SMBC Group's Japan Research Institute deployed an AI factory running Nemotron models, while Rakuten Bank is using the NVIDIA Agent Toolkit to build transaction foundation models across its 18 million accounts.
