How AI could help – or hurt – Asia’s energy future

ARTIFICIAL intelligence is one of the world’s fastest-growing sources of demand for electricity.

Between 2025 and 2030, data centres for AI are projected to double their share of global electricity consumption to around 3 per cent. That is a 30-fold increase since the start of the decade.

Currently, nearly two-thirds of the power that feeds AI comes from fossil fuels. Left unchecked, AI’s growth will lock in a new wave of carbon-emitting energy systems at exactly the moment when climate change and the global fuel crisis point to the urgent need to phase them out.

However, AI’s growing energy footprint is only one side of the story.

Can AI accelerate the energy transition?

AI may also prove to be a powerful tool for reducing energy waste, modernising energy systems and accelerating the transition to clean energy.

I see this evidence everywhere. In energy, transportation, agriculture and the built environment, it is supporting efforts to use less energy and facilitate the uptake of renewables.

In Asia, AI-powered irrigation systems are improving water efficiency and forecasting droughts and flooding. Across industry, the technology is pinpointing inefficiencies in production lines, shipping and logistics – reducing both costs and energy demand.

Yet the most transformative opportunity lies in making the world’s electricity grids modern and more efficient.

Under Global Energy Alliance’s India Grids of the Future Accelerator, the Indian state of Rajasthan has built a “digital twin” of Jaipur’s electricity grid, with around four million network assets mapped and digitised.

AI is now delivering real-time insights in seconds, spotting outages, predicting faults and modelling ways to integrate renewables and deploy battery storage.

Initial pilots suggest the digital twin could reduce annual operating costs by around US$15 million in Jaipur’s electricity network, including about US$1.2 million a year from improved transformer maintenance alone.

The programme could well improve energy reliability and clean energy supply for up to 18 million people.

In Asia, power utilities are waking up to the possibilities, harnessing AI to forecast energy demand, identify waste and optimise grid performance.

These changes are not “nice-to-have” benefits. The potential for carbon emissions reductions is striking.

Scientists estimate that deploying AI in global renewable energy systems could result in emission reductions to the tune of 1.8 gigatonnes of carbon dioxide a year by 2035 – 3.1 per cent of global emissions.

That is more than the annual carbon footprint of Japan, or indeed that of any country outside of China, the US, India and Russia.

The benefits extend well beyond carbon emissions.

Despite rapid progress, tens of millions of people in Asia still live in energy poverty, notes the International Energy Agency (IEA). By harnessing AI to make energy systems more efficient and cleaner, access to energy can increase without the heavy carbon price tag.

This approach is even more crucial in Africa, where the majority of the world’s energy-poor live – nearly 600 million people.

South-east Asia must manage the risks

But we must be clear-eyed about the risks. Asia, home to roughly 60 per cent of the world’s population, is in the middle of a data centre boom. Its share of global data centre capacity is projected to double to 40 per cent by 2030, said Moody’s.

Unless this rapid growth is matched by major investments in clean energy, AI risks driving a new cycle of carbon-intensive growth and deepening energy vulnerability among Asia’s emerging economies.

The IEA estimates that annual investments in South-east Asia’s grids and storage systems must rise from US$13 billion currently to US$50 billion in 2050 to deliver this transition.

The path forward for public policy is clear: Regional policymakers must strategically harness AI to strengthen energy security, expand access to electricity and reduce carbon emissions.

As an early mover, Singapore has already built a digital twin of its 27,000 kilometres of grid and other network assets. The city state is now integrating AI into daily field work and plans to use it for active grid control next.

In doing so, Singapore offers a case study of what a mature version of the Jaipur model could look like a decade from now.

Steering AI for public benefit will require strengthening Asia’s often still-nascent legislative and AI governance frameworks. Singapore has taken the regional lead by launching the world’s first Model AI governance framework and a Green Data Centre Roadmap mandating that new data centres source at least half their power from renewables.

Implementing stable, predictable and transparent regulations in the rest of Asia will help mobilise the enormous investments required to accelerate the adoption of clean energy, strengthen energy systems and drive economic opportunity.

The stakes are high. The size of AI’s climate footprint will depend on whether the carbon emitted to run it exceeds the carbon it helps avoid.

Countries that fail to align AI growth with clean energy risk rising costs, weaker productivity and greater exposure to volatile energy markets.

Those who pair AI with clean energy, grid modernisation and sound policies will create a lasting edge and future-proof both economic growth and social progress.

The writer is chief executive officer of Global Energy Alliance for People and Planet and former managing director-general at the Asian Development Bank