SINGAPORE, Jan 24, 2026, 21:19 SGT
- Singapore plans to pour over S$1 billion into public AI research by 2030.
- The plan focuses on basic research, practical industry applications, and expanding the talent pipeline from schools through to faculty levels.
- Officials highlighted the rising power and water expenses tied to running advanced AI as Singapore oversees the expansion of its data centres.
Singapore plans to pour over S$1 billion (around $780 million) into public AI research by 2030, the government announced Saturday. The funding aims to boost national AI capabilities and encourage wider adoption among businesses.
Funding arrives just as AI tools shift from demos to real-world use, while nations scramble to secure the computing power and talent needed to train and operate these models. The race remains dominated by U.S. and Chinese tech giants, with smaller economies looking to carve out niches.
Singapore is positioning this move as a research initiative rather than simply a procurement effort. The National AI Strategy 2.0, unveiled in 2023, aims to boost the count of AI practitioners to over 15,000—more than three times the current number. Officials are promoting the city-state as a prime location for top AI developers to establish teams.
The Ministry of Digital Development and Information plans to channel funding into priority fields like responsible, resource-efficient AI, while creating a talent pipeline from pre-university students up to faculty members. They’re also pushing for more practical tools that industries can actually use, not just academic papers.
The National AI Research and Development Plan, spanning from 2025 to 2030, was unveiled by Digital Development and Information Minister Josephine Teo during the Singapore AI Research Week 2026 gala dinner, the ministry confirmed. Teo cautioned that the computing costs are climbing. “For example, AI training and inference remain extremely resource-intensive. Their draw on energy and water cannot be ignored,” she stated. (Training involves the compute-heavy building of a model; inference uses that model to generate results once built.) Channelnewsasia
Teo said a large portion of the funds will be funneled into AI research centres of excellence, hosted by public research institutions and staffed with both local and international researchers. According to Business Times, these new centres will complement an existing network of over 60 AI centres of excellence, though they will be fewer and receive bigger budgets per centre.
The research agenda covers responsible AI—systems built with protections to prevent misuse—and efforts to cut AI’s dependence on data and computing power. Officials are also focusing on “general-purpose AI,” which refers to models capable of tackling a variety of tasks across multiple fields.
Applied research aims to bridge the gap between lab experiments and real-world deployment, the ministry announced. Efforts will include projects backed by national research and enterprise programmes. Authorities plan to collaborate with industry players like Changi Airport Group and Sembcorp to rapidly develop core AI engineering skills and pilot use-cases at scale, Business Times reported.
Singapore isn’t aiming to take on the global giants pushing the biggest “foundation models” headfirst — those massive AI systems trained on vast data sets and versatile across tasks. According to Business Times, the strategy is to develop core capabilities first, then customize models specifically for regional languages and security requirements.
That strategy mirrors efforts by AI Singapore, a national initiative supporting Sea-Lion, an open-source large language model tailored for Southeast Asian languages. Reuters noted that firms like Indonesia’s GoTo have already implemented the model.
Reuters reported that the updated Sea-Lion model, launched in October 2025, is based on Qwen, a foundation model from China’s Alibaba. This new version offers better support for Burmese, Filipino, Indonesian, Malay, Tamil, Thai, and Vietnamese. Separately, Singapore has allocated S$500 million for high-performance computing resources, the kind of powerful computing power needed to train and operate AI systems.
But money won’t erase physical barriers. Teo highlights AI’s heavy demand on electricity and water, along with the challenge of balancing data centre growth with Singapore’s net-zero goals. If computing power hits a bottleneck or businesses hesitate to adopt, turning research into real products could drag on.
Teo emphasized the importance of keeping the ecosystem both diverse and closely connected, avoiding silos. “We believe good outcomes will emerge out of a vibrant, diverse yet close-knit research ecosystem,” she said. Com
