How IMDA Shapes the Future of Singapore's Digital Economy
Few national bodies are tasked with simultaneously governing a digital economy, prototyping its frontier technologies and designing the policy frameworks that will govern them for decades.
Yet that is precisely the mandate of Singapore's Infocomm Media Development Authority (IMDA). Across quantum communications, enterprise AI, digital sustainability and tech policy, IMDA’s leadership team is working at a pace and depth that reflects the urgency they attach to the task.
Singapore’s scale is both a challenge and an advantage. Dr Lawrence Wee, Director of Business and Ecosystems at IMDA, puts it plainly: “Singapore hosts one of the region’s densest digital infrastructures. So there is what we call a digital paradox — the technology that we need for progression is also highly energy-intensive. Digital sustainability is very important for our future and it is the only way we can scale and ride on this AI revolution within the limits of our carbon and energy limits.”
That tension – between the demands of an advanced digital economy and the physical and environmental constraints of a small island state – runs through much of IMDA’s work. It also explains why the authority spans a broader remit than many comparable national bodies, bringing together engineering, policy, applied research and industry development under a unified strategic direction.
Building green software practice from the ground up
Lawrence’s work in digital sustainability is grounded in a practical, sector-facing approach. Rather than issuing broad directives, IMDA has developed resources that meet enterprises where they are, including the practical green software playbook, which breaks down sustainability into actionable technical steps.
“We launched the practical green software playbook for enterprises,” Lawrence explains. “We can break it down into three areas: tips to tell enterprises how to start measuring and monitoring energy consumption, tips around how to optimise the way enterprises can design software to reduce energy consumption and, in the area of AI, tips on how we can rightsize the different AI models to fit a specific use case.”
The rightsize principle has particular commercial relevance. Lawrence points to Gen AI as an area where enterprise decisions frequently carry hidden environmental and financial costs: "Enterprises can unlock greater value by matching AI models to their specific use cases. A good way would be to look at the specific use cases they have in store and then consider using much smaller language models and fine-tuning them on proprietary enterprise data. When they use a much smaller model fine-tuned on their data sets, the total cost will be a lot lower than using just the frontier models and the accuracy could even be better.”
IMDA has also aligned its approach with international frameworks. The authority adopted the Green Software Foundation’s Software Carbon Intensity specifications, enabling organisations to identify inefficiencies and baseline their carbon footprint across digital operations. Lawrence describes the approach as one of leading by example: IMDA’s own engineers are asked to implement green techniques within the software development process and the authority works directly with data centres to combine green software recommendations with operational changes at the infrastructure level.
Rapid prototyping as national infrastructure strategy
Where Lawrence focuses on sustainable adoption, Dr Clifton Phua, Director of Labs at IMDA, approaches the authority’s mandate from a slightly different angle: moving frontier technologies from research concept to deployable reality as rapidly as possible.
“At our Labs, my team operates like a tech SWAT team,” Clifton says. “We unbox frontier technologies from first principles and rapidly prototype them. Instead of just studying quantum computing or embodied AI, we actually build solutions that tackle real-world business problems today so that we can empower Singapore’s workforce with cutting-edge tools.”
The Labs function within IMDA’s BizTech Group is explicitly designed to avoid the pitfall of technology remaining in the lab while industry waits. Clifton describes a process built around minimum viable products and direct co-design with the businesses that will eventually use the tools: “Innovation doesn't happen in a vacuum. We actively co-design solutions with government regulators, end users of the technology, as well as working with tech companies. By collaborating across government, industry and research, we ensure our prototypes actually scale into ecosystems that deliver genuine productivity gains.”
The infrastructure layer is central to this ambition. Clifton is candid about the stakes: "Compute is the engine of enterprise AI. Without robust, scalable and sovereign computing infrastructure, Singapore’s ambitions in advanced AI will hit a hard ceiling. However, competitiveness isn’t just about raw power. It’s also about sustainable power. By investing in efficient AI compute, we will actually help enterprises lower their operational costs and at the same time meet their net-zero ESG targets.”
Quantum-safe networks and the long infrastructure cycle
While AI dominates much of the public conversation around digital transformation, Mr Wee-Sain Koh, Cluster Director of Engineering at IMDA, is focused on a slower-moving but equally significant technology transition: the shift to quantum-safe communications infrastructure.
Wee Sain’s path to quantum began with foundational work on Singapore's digital connectivity – the National Broadband Network, the Wireless@SG programme – before a visit to the Centre for Quantum Technologies in 2022 changed his perspective on what was genuinely deployable. “While many thought quantum was very nascent, they showed us commercial products with relatively stable performance,” he recalls. “We realised that Singapore would need to plan ahead to align with tech refresh cycles.”
The National Quantum Safe Network Plus (NQSN+) initiative is the result of that assessment. Its urgency is rooted in a straightforward problem: quantum computing will eventually be powerful enough to break today's encryption standards and the transition timeline is long. “Transitioning systems to become quantum-safe can take seven to 10 years,” Wee Sain explains.
NQSN+ began with Quantum Key Distribution but is evolving. “Operators are evolving their suites to include Post-Quantum Cryptography and other technologies to offer comprehensive solutions as the ecosystem matures,” Wee Sain notes.
His longer-term vision positions Singapore as a hub for quantum connectivity in the same way it has served as a hub for undersea cables and satellite communications. “Singapore has always been a connectivity hub – for subsea cables, satellites and data centres. I plan for Singapore to continue being a hub for quantum networks,” he says. “This includes building quantum key distribution satellites and evolving submarine cables to adopt fibre-based quantum networks and repeaters.”
Where engineering meets policy in digital resilience
Wee Sain’s role is distinctive in that it sits directly at the boundary between technical analysis and policy formation. The same expertise that informs engineering decisions feeds into the regulatory frameworks that govern how Singapore's critical infrastructure operators manage risk.
“Engineering analyses emerging trends to understand both benefits and technical risks,” he says. “For example, AI can improve network efficiency, but we must analyse the risks of fully autonomous infrastructure, such as explainability and new attack surfaces. This understanding helps set policy, such as deciding if a ‘human in the loop’ is required for final decisions.”
The same logic applies to quantum. Post-Quantum Cryptography and Quantum Key Distribution each carry different technical limitations and different levels of international acceptance and policy must account for both: “Policy must consider both limitations and potential to decide the path Singapore should take.”
Governing AI at scale across enterprise and industry
Arisa Siong, Director of Tech Policy at IMDA, approaches these questions from a governance and policy perspective built on years of working in telecoms regulation across Asia. Her focus is on ensuring that the frameworks IMDA develops are grounded in the practical realities of enterprise deployment rather than theoretical compliance – and that they account for the cultural and contextual dimensions that generic global models often miss.
One of the clearest examples concerns the limits of AI systems trained primarily on Western data. “A Western-trained Vision Language Model might identify a clock as a great gift, completely missing the profound cultural taboo associated with it in Chinese traditions,” Arisa explains. “To ensure AI systems reflect the full richness of global cultures, we are actively collaborating with global bodies like MLCommons – alongside tech leaders like Google and Microsoft – by developing and expanding comprehensive global safety benchmarks to evaluate models for multicultural nuances and complex vulnerabilities.”
Through the National Multimodal LLM Programme, IMDA is also building locally anchored alternatives. Models such as SEA-LION and MERaLiON are designed to reflect Southeast Asia’s linguistic and cultural range, ensuring that Singapore’s national identity and regional context are embedded in the foundational technology itself.
IMDA’s GPT-Legal initiative illustrates the practical impact of this approach. Rather than providing law firms with a generic AI assistant, the authority established a design partnership with the Singapore Academy of Law to build a tool calibrated specifically to Singapore’s legal context. The result was adoption by 70% of lawyers performing legal research in Singapore – a figure that speaks to what Arisa describes as the conditions for genuine enterprise scale: “Scale happens when you reduce manual bottlenecks and enable seamless innovation, empowering your workforce to focus on what matters most while AI handles the routine.”
Closing the governance gap across enterprise sizes
Arisa identifies the uneven distribution of AI governance capacity as one of the more pressing structural challenges in enterprise AI deployment. Large organisations can invest in dedicated governance boards and embed safety into design from the outset. Smaller companies often cannot.
“There are different starting points for how differently sized companies can manage these risks,” she says. “Large enterprises typically have the R&D resources to establish dedicated cross-functional governance boards and embed safety directly into their design. In contrast, smaller companies often lack these specialised resources, leaving them with a narrow scope of AI adoption and a heavy reliance on generic vendor assurances.”
Project Moonshot, IMDA’s AI testing toolkit, is designed to address this gap directly by giving companies of all sizes access to rigorous evaluation frameworks. The underlying philosophy is that responsible AI operations should not be contingent on organisational scale.
The coherence of IMDA’s approach across all four domains – sustainability, applied innovation, quantum resilience and AI governance – reflects a deliberate institutional design. The BizTech Group’s engineers build and test systems, the policy teams govern them and the ecosystem-facing work ensures adoption moves beyond proof of concept into durable industry transformation. As Arisa puts it: “We don’t just participate in the global tech revolution, but ensure our national identity, culture and aspirations are reflected in every innovation we shape.”
