NVIDIA says Japanese organizations build AI with NVIDIA Nemotron open models

NVIDIA said Japanese enterprises, startups, and research institutions are building AI models and applications with NVIDIA Nemotron open models, data, and libraries. The effort centers on customization for Japanese-language and locally oriented use cases, with deployments described across multiple organizations.

NVIDIA linked the adoption of open models to organizations’ ability to customize, deploy, and govern AI they control, and it cited workforce transition and an aging population as factors increasing demand for AI tailored to local needs. The release also described use cases spanning financial-document translation, enterprise knowledge applications, and telecom operations.

Institute of Science Tokyo developed the Swallow family of open foundation models using NVIDIA Nemotron datasets and the NVIDIA NeMo software stack for continual pretraining and post-training, aiming to enhance Japanese language and reasoning performance while keeping underlying English, math, and coding capabilities. SB Intuitions trained the Sarashina series using NVIDIA Nemotron, including NVIDIA NeMo RL and Megatron-LM libraries, and said SoftBank Corp. developed and deployed a large telco model using NVIDIA Nemotron for autonomous telecom network operations. Stockmark said its specialized Japanese-language document-understanding model is based on the NVIDIA Nemotron 3 Nano Omni model.

Other deployments described in the release included avatarin using NVIDIA Nemotron open models and NVIDIA NeMo for Japanese-language speech and reasoning for AI agents, with NVIDIA HGX B300 systems and NVIDIA Jetson powering related capabilities. ENEOS Holdings used NVIDIA Nemotron open models with NVIDIA AI-Q Blueprint and NVIDIA ALCHEMI NIM microservices for agentic AI workflows for energy and materials R&D, while NTT DATA used NVIDIA Nemotron-Personas-Japan to augment training data for its tsuzumi 2 model. Hitachi said it developed physical AI technologies using NVIDIA Nemotron and NVIDIA Cosmos open models within a multi-agent orchestration platform. Sakana AI said it collaborated with NVIDIA to integrate NVIDIA Nemotron into its Fugu model-orchestration platform for routing requests to models for different tasks, and NVIDIA stated Nemotron models are released with open weights, datasets, and recipes and can be used with NVIDIA NeMo for customization and deployment.

“Every nation and every company should own and control its intelligence infrastructure. Open models make that possible,” said Jensen Huang, founder and CEO of NVIDIA.

Provided by Globe Newswire on behalf of Nvidia. Click to read original content.