Artificial intelligence (AI) could add $53 billion to Singapore’s economy over the next 60 years, according to the Economic Opportunity Report 2025 by consultancy firm Public First, commissioned by Google. The largest gains are expected from productivity improvements, with AI freeing up nearly 21 working days annually for the average worker by automating routine tasks. These time savings could lead to wage increases exceeding 8%, as workers shift into higher-value roles.
While much attention has gone to training large language models, the more frequent and commercially critical task is AI inference, when AI models are used to generate outputs in real time. From fraud detection, real-time translation to customer support chatbots, inference underpins most enterprise AI applications in the real world.
Key enablers of AI inferencing include high-performance data centres, fast and low-latency connectivity, and energy-efficient compute environments. The good news is that Singapore has built a competitive edge in these areas, thanks to its government support and efforts by its infrastructure players like Keppel and Singapore Telecommunications (Singtel).
National initiatives
To capture the economic opportunities of AI, Singapore rolled out its National AI Strategy (NAIS) in 2019, laying the foundation for AI to transform key sectors such as transport, healthcare, education, and security. In 2023, the strategy was updated as NAIS 2.0, with a sharpened focus on “AI for the Public Good, for Singapore and the World”. The latest version outlines two core ambitions: developing peaks of excellence in AI capabilities and empowering local communities and businesses to use AI in a trusted and inclusive manner.
Complementing this strategy is the Enterprise Compute Initiative (ECI), which encourages companies to develop AI-driven solutions and build in-house AI capabilities. It does so by offering cloud computing credits, technical advisory, and change management support. For example, Google Cloud and Digital Industry Singapore’s AI Cloud Takeoff programme under the ECI seeks to establish AI Centres of Excellence in 300 Singapore-based companies within a year.
See also: AI agents expected to drive revenue by Apac CFOs, adoption rising in Southeast Asia
However, policy alone is not enough. For AI (especially inference) to work at scale, it needs a strong foundation of digital infrastructure to support the demand for low latency, high throughput, and dense compute power. This is where Singapore’s private sector enablers come in.
A strong AI inference backbone
According to Morgan Stanley Research, Singapore’s telecom and infrastructure players, particularly Singtel and Keppel, are uniquely positioned as AI enablers. Together, they are investing billions into building and powering next-generation data centres tailored for AI inference workloads.
See also: Traders are fleeing stocks feared to be under threat from AI
At the core of this digital infrastructure buildout is connectivity. Singapore is already one of Asia’s most connected nations with 26 subsea cable landings, which is the digital equivalent of expressways for cross-border data flow. These fibre-optic cables transmit data at near-light speed, enabling instant access to cloud platforms and real-time AI services. To meet rising demand, Singapore plans to double its subsea capacity over the next decade.
Several major cable projects are already underway. Keppel’s Bifrost Cable System will connect Singapore directly to the west coast of North America via Indonesia.
Meanwhile, a Singtel-led consortium and NEC Corporation are building a new 8,900km submarine cable system called the Asia United Gateway East (AUG East). Scheduled for completion by the third quarter of 2029, AUG East will connect Singapore to Japan, and also touch Brunei, Indonesia, Malaysia, the Philippines, South Korea and Taiwan. It brings on board a heavyweight consortium including Amazon Web Services, Microsoft, Aeteria Networks, Chunghwa Telecom, Dreamline, Globe Telecom, Telekom Malaysia and Unified National Networks.
Domestically, the government aims to support broadband speeds up to 10Gbps within five years to ensure that businesses and consumers can seamlessly access increasingly data-intensive services.
On the compute side, Singapore is home to around 70 data centres that collectively offer 1.4 gigawatts (GW) of operational IT capacity, placing it at the top of the Asean region. However, growth is expected to moderate due to land and power constraints, with capacity expanding at 3% annually, down from 5% during the previous cycle.
Still, the outlook remains robust. “We expect at least 500 megawatts (MW) of incremental data centres in Singapore by 2030, with potential for 1.5GW as additional power generation assets are added,” wrote Morgan Stanley Research analysts in a June 23 report.
Future data centres are being designed specifically for AI. A prime example is DC Tuas by Singtel’s regional data centre arm Nxera, which will begin operations in 2026. The facility will deliver 58MW of IT load, with liquid cooling systems supporting up to 150 kilowatts per rack, and a submarine cable landing station integrated directly into the site for regional and global hyper-connectivity.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
“It will be one of the largest liquid-cooled AI data centres in Asia, with the highest power and water efficiency,” says Bill Chang, CEO of Singtel’s Digital InfraCo and Nxera, at a July 25 media briefing. Similar to other Nxera data centres, DC Tuas will also feature advanced telemetry systems and a digital platform to monitor all critical systems and extract real-time operational insights for performance optimisation. “Instead of referring to them as data centres, I would liken them to Centres of Intelligence that accelerate innovation and transform enterprises, industries, and economies through the power of AI,” Chang adds.
Supporting this compute expansion is a growing need for a reliable and scalable energy supply. Keppel is addressing this by planning a 600MW hydrogen-ready gas-fired power plant that is designed to meet the increasing energy demands of AI infrastructure and digital services.
Morgan Stanley Research expects incremental cable networks and new AI data centres to bring in $20 billion in investments and lay the foundation for Singapore’s next phase of economic growth.
Johor as Singapore’s digital extension
Singapore’s land and energy constraints are driving spillover data centre investments into Johor, Malaysia, which benefits from its proximity and lower operating costs. The Johor-Singapore Special Economic Zone (JS-SEZ) reinforces this trend, positioning Johor as a natural extension of Singapore’s digital infrastructure as regional demand for AI inference and data processing continues to grow.
This has already shown results. Malaysia’s Bernama reported on July 29 that Johor has approved 42 data centre construction projects in the second quarter of this year.
Moreover, Singapore-based digital infrastructure companies are increasingly setting up operations in Johor. Princeton Digital Group launched its 150MW greenfield data centre campus in Sedenak Tech Park in July last year. Its chairman and CEO, Rangu Salgame, sees that campus playing a vital role in training foundation models.
“During training, the large-scale foundation models that have emerged over the last two years are not transferring [or handling] end-user data in real-time. That means they can be located in regions like Johor, where land and power are more abundant, slightly away from end-users but still close enough to tap into the global subsea cable network,” says Salgame at the recent Fortune Brainstorm AI conference in Singapore.
Meanwhile, Nxera has partnered with Telekom Malaysia to jointly develop a 64MW AI-ready, sustainable data centre campus in Iskandar Puteri. Set to begin operations in 2026, the facility is designed to scale up to 200MW as demand grows. Its location (which is just 16km away from Singapore) enables enterprises to expand seamlessly from Singapore while maintaining low-latency performance.
Still, Singapore’s data centres continue to offer strategic advantages, particularly for companies that must keep data close to their end users or comply with data sovereignty requirements, says Singtel’s Chang. This includes multinational corporations headquartered in Singapore and cloud providers serving customers that require data to reside or be stored locally.
He also highlights Singapore’s position as a digital and business hub, supported by “abundant connectivity such as submarine cable landings, and strong domestic infrastructure like 5G networks”. These factors are critical when companies manage workloads across multiple countries.
“When these companies think about their load balancing [or distributing computing tasks across multiple locations to reduce latency], they don’t just think about their zones in Singapore — they think about multi-country support. So, those elements work well for Singapore’s data centres, [especially for AI inferencing],” adds Chang.
As AI adoption moves from experimentation to enterprise-wide deployment, the real competitive advantage lies not only in algorithms but in the infrastructure that powers them. With its robust connectivity, AI-ready data centres, and strong policy foundation, Singapore is well-positioned as the regional hub for AI inference.
