Building a sustainable AI-driven digital nation

Sustainable artificial intelligence (AI) sounds like a contradiction, as AI is a major energy guzzler with power hungry data centres being a major requirement.

According to the International Energy Agency (IEA), a typical AI-focused data centre consumes as much electricity as 100,000 households, and this would only continue to increase as demand for AI grows.

But there’s another side to the coin, as AI could also help to achieve more sustainable energy management.

This was the paradox that experts at the GovInsider Festival of Innovation 2026 discussed during the panel session Building a Climate-Resilient Digital Nation.

The panel included energy, academia, and international organisation leaders who argued that sustainable AI is not just aspirational, but achievable.  

Three pillars for achieving sustainable digitisation 

The speakers noted that achieving sustainable AI required three structural elements: policies that reward efficiency, energy-efficient computing infrastructure, and a cleaner electricity supply.

Singapore’s Energy Market Authority (EMA) Chief Data Officer, Chua Shen Hwee, noted that AI was a tool that could be used to optimise energy management and forecast solar energy outputs.

Since solar energy output depends on cloud coverage, the power system control officials need to prepare sufficient thermal generation to ensure a stable electricity supply during cloudy and rainy days.

“While AI contributes to the demand, it also helps us to optimise our systems by monitoring and forecasting if there’s cloud cover that we need to consider in order to step up in terms of thermal generation,” explained Chua.

In this case, AI stood as a lever to enhance operations and energy management as Singapore moves to a diversified energy mix to reduce carbon emissions from power generation. 

Chua added that on a longer-term basis, EMA also leverages AI to consider climate change impacts such as increase in temperatures to project future temperatures and proactively ensure that Singapore can secure energy supply for the years to come. 

Regulations for progress 

Emphasising that regulations do not stifle innovation, the speakers stressed on the need to regulate outcomes, rather than AI itself. 

Forward-looking policies that set clear sustainability and energy efficiency standards can help shape industry behaviour without slowing down innovation, noted Chua.

She shared the examples of the Infocomm Media Development Authority (IMDA)’s green data centre standards that set the policies and standards for data centre operators to achieve energy efficiency, and the frameworks that outline clear intent on greening IT equipment efficiency.

UNDP’s Regional Digital & AI Expert for Asia Pacific, Juan Kanggrawan, added that forward-looking policies could be a catalyst in Southeast Asia given the context of the region. 

“We can see the ASEAN stance or policies about energy and AI and then see how we can relate that to country-level policy, then deeper in the provincial or even city level,” he said, adding that countries can navigate these policies in parallel and learn mutually. 

“If Singapore started somewhere, we [Indonesia] can learn from that. If other countries started somewhere else, we could learn from them as well. Even in the context of policy formulation and execution, it can be dynamic as well.” 

The ASEAN power grid (APG), which sought to connect the electricity networks of the ASEAN member countries to enable cross-border power trading and ensure a reliable energy supply, recently saw the signing of an enhanced memorandum of understanding (MoU) on standardising rules and regulations for cross-border electricity trade.

GovInsider previously reported how ASEAN leaders were calling for join action in developing regional guidelines for AI growth taking into account the different levels of readiness and needs of each member country. 

Clarity on all fronts 

Clear intent about both technology use and sustainability were the main points that experts discussed.

Chua noted that this helped to ensure that sustainability efforts remained focused on objectives and were therefore more effective. 

Organisations should also resist adopting technology simply because it's trendy, noted Kanggrawan. 

“Before implementing AI or any new technology, clearly define organisational strategy, specific needs, and what you want to achieve in the next three to 10 years. Start small, low hanging fruits, not so complex, to then iterate and see the impact,” he noted.

NUS-ISS Chief of Data Science & Digital Sustainability, Clara Lee, added that understanding the sustainability risks could also help organisations better prepare for strategies and actions to mitigate risks and unlock returns, be it profits for private sector or national goals for public sector.

With this clarity, many successful companies demonstrated that sustainability and business objectives were not in conflict, noted Lee.

Integrating sustainability throughout their operations, from product design to supply chain, they were able to maintain business success and attract talent.

She shared a simple example: “if you look at their websites, they're all dark mode, which is more energy-efficient than the light mode. That’s how you see that sustainability is everywhere in their design”. 

There’s no tomorrow without sustainability 

As governments around the world embrace digital transformation to improve citizen services and operations, sustainability must be at the core of the conversation.

Recognising AI’s environmental impact meant taking a critical and holistic approach to understanding where it adds value, so that organisations can use it more mindfully.

“Everybody now learns that the more AI we do, the more technology consumption it takes. I would like to see that moving forward we do not just measure its financial impact, but also environmental impact and [make it] a forefront topic,” said Lee. 

You may revisit this session here!