Japanese enterprises, startups, and research institutions are building industry-specialized AI models and applications with NVIDIA Nemotron open models, data, and libraries. This development accelerates the creation of AI tailored to Japan’s language, industries, and workforce. The adoption of open models allows organizations to customize, deploy, and govern AI they control, which is increasingly important as Japan addresses an aging population and workforce transition.

Institute of Science Tokyo developed its Swallow family of open foundation models using NVIDIA Nemotron datasets and the NVIDIA NeMo software stack. The Swallow models enhance Japanese language and reasoning performance while preserving core English, math, and coding capabilities. Enterprises are deploying these models for specialized use cases, including financial-document translation and asset-management report generation.

SB Intuitions Corp., the generative AI research subsidiary of SoftBank Corp., trained its Sarashina series of models using NVIDIA Nemotron, including the NVIDIA NeMo RL and Megatron-LM libraries. The Sarashina3 mini model has been selected by Japan’s Digital Agency for specialized AI use cases. Additionally, SoftBank Corp. has developed a large telco model using NVIDIA Nemotron to enable autonomous telecom network operations.

Stockmark released a specialized Japanese-language document-understanding model based on the NVIDIA Nemotron 3 Nano Omni model. The company is developing enterprise knowledge applications using NVIDIA NeMo Retriever and the Nemotron-Personas-Japan dataset. These applications serve customers in manufacturing, energy, and chemical industries through Japan’s Generative AI Accelerator Challenge national project.

Several Japanese enterprises are leveraging NVIDIA Nemotron to modernize services and improve productivity. Avatarin is using the models to develop Japanese-language speech and reasoning capabilities for enterprise AI agents. ENEOS Holdings is advancing agentic AI workflows for energy and materials R&D using NVIDIA Nemotron with the NVIDIA AI-Q Blueprint and NVIDIA ALCHEMI NIM microservices. NTT DATA augmented training data for its tsuzumi 2 model using NVIDIA Nemotron-Personas-Japan to improve question-answering accuracy. Hitachi is developing physical AI technologies using NVIDIA Nemotron and NVIDIA Cosmos open models to address operational challenges.

Sakana AI is integrating NVIDIA Nemotron into its Fugu model-orchestration platform. This integration expands the range of AI models Fugu can intelligently orchestrate to dynamically select the best model for each task. By routing requests to the most suitable model, Fugu helps developers balance accuracy, performance, and cost across multiple open and proprietary AI models.

NVIDIA Nemotron models are released with open weights, datasets, and recipes, providing transparency and control for domain-specific customization. Developers can use NVIDIA NeMo to customize and deploy models in environments that meet regulatory and data localization requirements. The models are available on Hugging Face, ModelScope, OpenRouter, and build.nvidia.com as NVIDIA NIM microservices, as well as through NVIDIA Cloud Partners and cloud service providers.

NVIDIA announced on July 15, 2026, that Japan’s physical AI leaders are building on the NVIDIA Cosmos, NVIDIA Isaac, NVIDIA Metropolis and NVIDIA Jetson platforms to accelerate the deployment of intelligent machines across manufacturing, mobility, infrastructure and robotics. The company introduced Cosmos 3 Edge, a new addition to the NVIDIA Cosmos 3 open world model family, which brings frontier capabilities to NVIDIA Jetson Thor platforms to help embodied systems see, reason in real time and predict robot actions locally.

NVIDIA Cosmos 3 Edge is a 4-billion-parameter model built on NVIDIA Nemotron that helps robots and vision AI agents understand their surroundings, reason in real time and generate robot actions on NVIDIA edge computers. Using the open NVIDIA Cosmos framework, developers can adapt the model for specific robots, vehicles, sensors and environments in about a day. The model is lightweight enough to run on edge GPUs and can be deployed across NVIDIA RTX GPUs, NVIDIA DGX systems and NVIDIA Jetson, including the newly announced T2000 and T3000 modules.

To further accelerate the development of vision AI agents, NVIDIA is also announcing new NVIDIA Metropolis libraries and skills that help developers use coding agents to build, train and operate video intelligence systems with Cosmos at least 6x faster. NVIDIA is expanding the NVIDIA Cosmos Coalition to Japan, bringing together world model builders, AI developers and physical AI leaders to advance open world models with Cosmos technologies.

Japan’s physical AI ecosystem leaders including AIRoA, classmethod, Enactic, FANUC, Fujitsu, GROOVE X, Hitachi, Honda R&D, Kawasaki Heavy Industries, Kubota, Mitsui & Co., Mitsubishi Corp., Mujin, NEC, Preferred Networks, SoftBank Corp., Sony Group Corporation, Telexistence, TIER IV, TRON K.K., Turing and Yaskawa Electric intend to join the coalition. Coalition members can contribute to and build on the NVIDIA Cosmos platform, which includes open models, data curation libraries, datasets and frameworks.

Fujitsu is exploring business opportunities in physical AI with FANUC, Yaskawa Electric and Kawasaki Heavy Industries. Led by Fujitsu, the initiative aims to build a collaborative control platform integrating NVIDIA’s physical AI stack to bridge digital and physical operations across all industrial sectors. Built with Cosmos world foundation models, the open Isaac robotics development platform, NVIDIA Omniverse NuRec libraries and the Newton physics engine, the platform will support AI model development, digital twins, robot learning, simulation-to-real workflows and pre-deployment validation.

NEC, Hitachi, OMRON and Preferred Networks are using NVIDIA Cosmos and NVIDIA physical AI technologies to advance world models, industrial AI and physical AI R&D. SoftBank Corp. is developing a physical AI development platform built on NVIDIA Cosmos, NVIDIA Omniverse and NVIDIA Isaac Sim. The company is also advancing AI-RAN initiatives using NVIDIA AI Aerial with the aim of delivering intelligent connectivity for billions of physical AI devices. Mujin is exploring NVIDIA Cosmos for autonomous robotics and intelligent industrial automation powered by MujinOS, while TRON K.K. is developing manufacturing data workflows for task-specific physical AI models in assembly, picking, inspection and material handling, as well as factory 3D digitization workflows.

Kawasaki Heavy Industries is applying NVIDIA physical AI technologies across healthcare, shipbuilding, transportation, aerospace and energy; Kubota is exploring Cosmos-based physical AI for autonomous agriculture and smart farming. Enactic is fine-tuning the NVIDIA Isaac GR00T open model for elder-care semi-humanoid robots; GROOVE X is building Jetson-powered companion robots, LOVOT; and Telexistence is applying Isaac and exploring Cosmos for retail automation. Japan’s industry leaders are also using NVIDIA Metropolis to bring Cosmos-powered vision AI agents into physical operations: Hitachi for smart-building operations, OMRON for automated inspection and Shimizu Corporation for construction safety.