Nvidia, a leader in the semiconductor industry, plans to support the development of artificial intelligence (AI) tailored to the Japanese language. In a recent announcement, as reported by The Japan Times, Nvidia said it will expand AI development services in Japan, focusing on large language models (LLMs) trained on Japanese data. This initiative aligns with broader efforts to enhance the country’s national economic security by fostering AI that is independent of foreign data sources.

The expansion includes collaborations with notable Japanese institutions such as the Tokyo Institute of Technology and Rakuten Group. These partnerships aim to build AI models that are culturally and linguistically adapted to the Japanese context. According to Nvidia CEO Jensen Huang, this move is essential for Japan to maintain control over its data and AI development. “There’s no reason to allow some other third party to harvest that data [of Japan], create an AI, and then import it back to Japan,” Huang emphasized earlier this year.

This initiative is part of Nvidia’s broader strategy to support sovereign AI, which involves developing AI systems using local computing infrastructure, data, and expertise. This approach aims to ensure that AI applications are closely aligned with regional values, laws, and cultural norms. To that end, the company has introduced four new NIMs (Nvidia Inference Microservices) aimed at accelerating the deployment of generative AI applications in both Japan and Taiwan, according to a blog post. These microservices include models such as Llama-3-Swallow-70B, which is fine-tuned on Japanese language data, and Llama-3-Taiwan-70B, tailored for Mandarin.

The Tokyo Institute of Technology has been instrumental in fine-tuning these models, emphasizing the importance of developing AI that adheres to local cultural norms. “LLMs are not mechanical tools that provide the same benefit for everyone. They are intellectual tools that interact with human culture and creativity,” said Rio Yokota, a professor at the Global Scientific Information and Computing Center at the Tokyo Institute of Technology.

As generative AI software revenue in the Asia-Pacific region is expected to surge from $5 billion this year to $48 billion by 2030, Nvidia’s initiatives in Japan and Taiwan represent a significant step toward ensuring that these regions remain competitive in the global AI landscape.