Nvidia says Sea Limited, the Singapore-based parent of Shopee, Garena and digital financial services arm Monee, has become the first company in ASEAN to adopt its new Vera Rubin platform, one of a string of Southeast Asia deployments the company detailed at AI Day Singapore.
The event ran from 22 to 23 September at Raffles City Convention Centre, offering hands-on training and expert-led sessions on AI and high-performance computing. Alongside Sea’s announcement, Nvidia laid out how government agencies, enterprises and developers across the region are putting its Nemotron open models, NeMo tools and Cosmos world models to work.
Sea Limited scales AI across its platforms
Sea plans to use the Vera Rubin platform to scale AI capabilities across Shopee, Monee and Garena. On Shopee, AI already helps sellers create product listings, improves product discovery, and supports small and medium-sized businesses in making decisions about their operations. Monee applies AI to fraud detection and credit risk assessment, while Garena uses it to enhance gameplay and build safer online spaces for players.
Singapore’s public agencies turn to Nemotron
Several Singapore public-sector organisations are working with Nvidia’s models. The Home Team Science and Technology Agency is researching Nemotron 3 Super and Nemotron 3 Nano Omni for public safety applications. NCS is advancing agentic AI adoption using Nemotron models together with an Nvidia Blueprint for video search and summarisation, and is also developing humanoid robotics applications. ST Engineering is using Nvidia’s NeMo tools and cuOpt software to build an AI Studio platform and agentic solutions for its Marine MRO business.
AI Singapore is expanding its SEA-LION family of models using Nemotron open models and NeMo tools, aimed at better supporting Southeast Asian languages and cultures. Separately, Hummingbird Bioscience is working with LynxKite to build an explainable Toxicity Knowledge Graph using Nemotron 3.5 Lightning and NeMo Retriever, intended to accelerate drug discovery.
Open models take root across the region
Nvidia also pointed to a wider set of deployments across Southeast Asia:
- Malaysia: YTL AI Labs is fine-tuning Nemotron models to power enterprise and citizen services, while ITMAX is using Nvidia Cosmos for video search and summarisation to manage city traffic operations.
- Vietnam: Viettel AI has fine-tuned Nemotron 3 Super for Vietnamese-language applications, where it topped the VMLU benchmark and the company’s internal product benchmark; it plans to use the model for a legal AI agent serving employees and customers. FPT Smart Cloud co-developed the Nemotron-Personas-Vietnam dataset, grounded in Vietnamese demographic and cultural data.
- Thailand: The ThaiLLM collaboration and iApp Technology adapted Nemotron 3 Nano with the OpenThai 2.0 Legal framework to build Thanoy, a legal-assistant chatbot that Nvidia says serves around 43,000 users. AS-TECH is applying Nvidia Cosmos to manage passenger flow at airports.
- Brunei: Antrique has built an AI platform aimed at improving productivity in the food sector.
GPU access for regional developers
Nvidia also worked with Bitdeer AI to co-host an Open Models AI Codefest, with Bitdeer providing GPU cloud infrastructure for developers building on the open models showcased at the event.
