At NVIDIA AI Day Singapore, NVIDIA announced major regional expansions of its AI infrastructure, open foundation models, and developer frameworks across Southeast Asia. The initiatives focus on enabling governments and enterprises to transition sovereign and localized AI projects from pilot phases to production-scale operations.

In Singapore, the Home Team Science and Technology Agency (HTX) is leveraging NVIDIA Nemotron 3 Super and Nemotron 3 Nano Omni models for public safety and multimodal reasoning. Tech service providers NCS and ST Engineering are integrating Nemotron, NVIDIA Blueprints, and cuOpt software to deploy agentic video analytics, physical AI for robotics, and logistics optimization across various commercial and defense operations.

Regional adoption spans several neighboring nations. YTL AI Labs in Malaysia, Viettel AI in Vietnam, and the ThaiLLM Collaboration in Thailand are actively fine-tuning open Nemotron models for local language services, legal chatbots, and civic infrastructure. Additionally, organizations in Malaysia and Thailand are deploying NVIDIA Cosmos world models to manage municipal traffic and airport passenger flows.

Why it matters

  • Enterprise operators in Southeast Asia gain localized, open-weight foundation models tailored for specific regional languages and compliance rules.

  • NVIDIA is solidifying its hardware dominance by embedding its software stack (NeMo, cuOpt, Cosmos) into national sovereign AI strategies.

  • Physical AI and agentic video applications are moving rapidly into real-world civic and industrial production across APAC.

Source: blogs.nvidia.com