NVIDIA has launched a suite of NIM microservices tailored for generative AI (GenAI) applications in Japan and Taiwan. As countries worldwide push for AI systems that align with local values and regulations, these new offerings enable developers to create and deploy high-performance AI applications sovereign AI capabilities in these regions.
Demand for GenAI in the Asia-Pacific region is rapidly growing, with software revenue expected to surge from US$5 billion this year to US$48 billion by 2030. NVIDIA’s NIM microservices are set to play a crucial role in this growth by supporting community models that enhance user interactions through a deep understanding of local languages and cultural contexts.
In Japan, the introduction of the Llama-3-Swallow-70B model, trained on Japanese data, is a pivotal development. This model is designed to understand and reflect Japan’s cultural and linguistic nuances, offering superior performance in tasks such as language translation, legal queries, and question-answering.
Tokyo Institute of Technology has fine-tuned this model to ensure that it adheres to Japanese cultural norms, enabling its deployment across various industries. For example, AI company Preferred Networks is leveraging this model to develop healthcare-specific applications, demonstrating its potential to transform industry practices.
In Taiwan, the Llama-3-Taiwan-70B model, trained on Mandarin data, is being utilised by leading institutions such as Chang Gung Memorial Hospital. This model helps streamline medical workflows by providing context-appropriate guidance in nuanced medical language to improve patient care. Pegatron and Chang Chun Group are among the Taiwanese companies that are adopting this model to enhance their operations, showcasing its versatility across different sectors.
NVIDIA’s NIM microservices facilitate the deployment of these AI models by providing a robust framework that simplifies the integration process. Available through NVIDIA AI Enterprise, these microservices are optimised for performance, offering up to five times higher throughput than base models and reducing latency for better user experiences. They offer an attractive option for businesses, government agencies and universities looking to harness the power of AI.
The launch of NIM microservices in Japan and Taiwan underscores the growing importance of sovereign AI infrastructure. Singapore, the United Arab Emirates, South Korea, and India are among other Asia-Pacific nations that are investing in this area.
Entelechy Asia 