NVIDIA has unveiled new artificial intelligence projects in Singapore, developed together with companies, public agencies, and research centers across Southeast Asia. Among the announcements, Sea Limited’s adoption of the NVIDIA Vera Rubin platform stands out, making it, according to NVIDIA, the first company in the ASEAN region to use this platform to expand its AI capabilities.
NVIDIA’s Singapore AI push in 20 seconds
- NVIDIA is bringing together AI projects from public agencies, companies, and research centers in Singapore.
- The focus is on Nemotron models, edge computing, and the Vera Rubin platform.
- Sea Limited will be the first ASEAN company to adopt NVIDIA Vera Rubin, according to NVIDIA.
- Nemotron models are being adapted to the languages and sectors of countries such as Vietnam, Thailand, and Malaysia.
- Use cases span public services, security, healthcare, mobility, e-commerce, and financial services.
NVIDIA AI Day Singapore, held on September 22 and 23 at the Raffles City Convention Centre, serves as a showcase for several projects the company is developing with partners across the region. The strategy combines compute infrastructure, open models, developer tools, and AI applications tailored to local languages and needs.
The company is paying particular attention to moving pilot projects into production deployments at public agencies. To that end, NVIDIA is working with partners in different countries on areas such as government services, citizen support, critical infrastructure, public safety, and support for startups, developers, and researchers.
Nemotron adapts to the region’s languages and needs
One of the main threads of the announcement centers on NVIDIA Nemotron, a family of open models that various Southeast Asian organizations are using to build applications tailored to their own markets.
In Singapore, HTX, the Home Team’s science and technology agency, is researching the use of Nemotron 3 Super and Nemotron 3 Nano Omni for public-safety applications. According to NVIDIA, Nemotron Super can be used for complex reasoning tasks and agentic workflows, while Nano Omni combines vision, audio, and language capabilities.
NCS, for its part, is working on adopting agent-based AI in both enterprise and public-sector settings. Its work includes Nemotron models and NVIDIA’s video search and summarization Blueprint, along with research into physical AI and humanoid robotics.
ST Engineering is also using NVIDIA NeMo tools and NVIDIA cuOpt software in developing its AI Studio platform. One of the areas mentioned is maintenance, repair, and overhaul of maritime assets.
Adapting to each country’s language characteristics is another thread of the project. In Malaysia, YTL AI Labs is fine-tuning Nemotron models for enterprise and citizen services. In Vietnam, Viettel AI is working with these models for Vietnamese-language applications.
In Thailand, the Big Data Institute and iApp Technology, as members of the ThaiLLM Collaboration, are exploring Nemotron as a foundation model. iApp is fine-tuning Nemotron 3 Nano with Thai legal data for OpenThai 2.0 Legal. The model has been released as open source for the Thai developer community and underpins Thanoy, a legal assistant that, according to NVIDIA, has around 43,000 users.
These models are also being used in smart-city applications. NVIDIA says its Cosmos open world models and its video search and summarization system are being used in projects in Malaysia and Thailand. ITMAX uses Cosmos alongside VSS for urban traffic operations, while AS-TECH uses the same combination to manage passenger flow at airports.
Sea Limited adopts NVIDIA Vera Rubin
The announcement with the biggest business footprint that NVIDIA presented in Singapore involves Sea Limited, the country’s own technology group and the owner of platforms such as Shopee, Garena, and Monee.
NVIDIA says Sea is the first company in the ASEAN region to adopt the NVIDIA Vera Rubin platform. The tech company will use this infrastructure to expand its capacity for developing and deploying AI models and agents.
Sea already uses AI across different parts of its business. At Shopee, the technology helps sellers create product listings, improve product discovery, and get insights for business decisions.
At Monee, its digital financial services arm, AI is used in areas such as fraud detection and credit-risk assessment. At Garena, its digital entertainment and gaming division, it’s used to improve various aspects of the gaming experience.
The addition of Vera Rubin is framed as an expansion of that computing infrastructure. NVIDIA says it will let Sea develop and deploy models and agents at a larger scale, though the announcement doesn’t specify how many systems will be installed, how much compute capacity has been contracted, or a specific timeline for going live.
The move is part of a broader NVIDIA strategy to push models and tools tailored to Asian markets. The company is positioning Nemotron as a foundation that organizations can customize and control for their own languages, sectors, and needs.
Beyond the models, NVIDIA is also making datasets and developer tools available. These include synthetic datasets designed to represent the region’s populations, languages, and professional profiles.
In Singapore, AI Singapore is expanding its SEA-LION family of models using NVIDIA Nemotron open models and NVIDIA NeMo tools. SEA-LION is designed specifically for Southeast Asian languages and cultures.
Other projects are more sector-specific. Hummingbird Bioscience and LynxKite are building a toxicity knowledge graph based on Nemotron 3.5 Lightning and NeMo Retriever, combined with in silico simulations. The stated goal is to bring together public and proprietary data from different formats to support drug discovery and development research.
The activity also reaches the developer community. Bitdeer AI, together with NVIDIA, organized the Open Models AI Codefest and provided GPU cloud infrastructure so developers across the region could work with Nemotron models, datasets, and training recipes.
In Vietnam, FPT Smart Cloud has developed, together with NVIDIA, Nemotron-Personas-Vietnam, an open dataset built on Vietnamese demographic and cultural information. Viettel AI, meanwhile, is fine-tuning Nemotron 3 Super for the Vietnamese language and agent-based applications.
The set of announcements presented in Singapore shows a strategy that combines compute infrastructure, open models, and market-specific applications. The projects are still at different stages, ranging from research and model fine-tuning to applications that already have users.
For NVIDIA, the goal is for these capabilities to let governments, companies, and developers build AI systems tailored to each country’s languages, sectors, and conditions. Sea Limited’s adoption of Vera Rubin adds another piece of infrastructure to a region where tech companies are expanding their use of artificial intelligence in e-commerce, financial services, entertainment, and public services.
FAQ
What has NVIDIA announced in Singapore?
NVIDIA has unveiled artificial intelligence projects developed with public agencies, companies, and research centers across Southeast Asia, alongside the announcement that Sea Limited has adopted Vera Rubin.
What is NVIDIA Nemotron?
Nemotron is a family of open AI models that various organizations in the region are adapting to specific languages, sectors, and applications.
Which ASEAN company has adopted NVIDIA Vera Rubin?
According to NVIDIA, Sea Limited is the first company in the ASEAN region to adopt the Vera Rubin platform.
What does Sea Limited use artificial intelligence for?
Sea uses AI across businesses such as Shopee, Monee, and Garena, with applications including product content creation, fraud detection, credit-risk assessment, and improvements to gaming experiences.
