NVIDIA proposed four areas of AI cooperation to Vietnam, including helping the country build its own large language model and expand its computing capacity, said Vikram Rao, NVIDIA’s vice-president for ASEAN.

In a statement on Thursday, the ministry noted the executive and Vietnamese Minister of Science and Technology Vu Hai Quan had a discussion on AI cooperation. NVIDIA said Vietnam had made progress in building its AI ecosystem, while NVIDIA has been already working on AI infrastructure with Vietnamese companies, including FPT, Viettel, and GreenNode.

Under the first proposal, NVIDIA offers to help Vietnam develop a large language model (LLM) that the country would own. The cooperation can feature NVIDIA’s open-source models, including its Nemotron family, tools, and data, in line with Vietnam’s own needs.

Regarding AI infrastructure and computing capacity, NVIDIA said Vietnam currently had about 10,000-15,000 graphics processing units (GPU). The figure is lower than 50,000-60,000 in South Korea, while South Korean plans to expand to 200,000-250,000 over the next five years. NVIDIA suggested Vietnam draw up a national plan to build AI computing capacity for government agencies, businesses, universities, and research.

The third proposal covers talent, startups, research, and education, such as expanding NVIDIA’s Deep Learning Institute training programs in Vietnam to a national scale and supporting Vietnamese AI startups, researchers, and students.

The fourth collaboration is to connect Vietnam’s AI sector to global markets, initially within Southeast Asia and later internationally.

Welcoming the initiatives, Minister Quan asked ministry units to work with NVIDIA to make detailed joint action plans.

NVIDIA aims to build Hanoi’s R&D center into world-class AI engineering hub