NVIDIA has released the Nemotron-Personas-Vietnam dataset in partnership with FPT, consisting of 900,000 synthetic personas tailored to Vietnam’s language, culture, and workforce. This initiative supports NVIDIA’s broader goal of promoting sovereign AI, which enables countries to create culturally aligned AI systems rather than relying on generic global models. The dataset is open-source and commercially available on Hugging Face, encouraging collaboration in localized AI development.

FPT: FPT is a leading Vietnamese technology services and digital transformation company with extensive operations in software, AI, and IT infrastructure. It collaborated directly with NVIDIA to produce and release the Nemotron-Personas-Vietnam dataset. The partnership focuses on building AI tools that better understand Vietnamese language, culture, and workforce realities.
NVIDIA: NVIDIA develops graphics processing units, AI accelerators, and software platforms that power data centers and generative AI applications worldwide. The company is actively advancing sovereign AI initiatives to help nations create localized models that respect cultural and regulatory contexts. In this news, NVIDIA partnered with FPT to release a Vietnam-specific dataset supporting those goals.
Hugging Face: Hugging Face operates a major platform for hosting, sharing, and deploying machine learning models and datasets. It serves as the distribution channel for the newly released Nemotron-Personas-Vietnam dataset. This enables global developers to access and build upon the Vietnam-specific AI resources.
Nemotron-Personas-Vietnam dataset: The Nemotron-Personas-Vietnam dataset consists of 900,000 synthetic personas tailored to Vietnam’s language, culture, and workforce characteristics. It was jointly developed and released by NVIDIA and FPT as an open-source resource. The dataset is hosted on Hugging Face for both open-source and commercial use.

Sovereign AI: NVIDIA continues to promote sovereign AI frameworks that allow countries to develop independent, culturally aligned AI systems rather than depending on generic global models.
Dataset Release: The Nemotron-Personas-Vietnam dataset is openly available on Hugging Face to encourage collaboration on localized AI development.