MUCH of the global debate on AI and sustainability fixates on how much electricity AI will require. In Asia-Pacific (APAC), the question is whether the region can scale AI infrastructure without undermining energy affordability, grid resilience and decarbonization goals—while simultaneously deploying AI to make energy systems smarter.
This AI-energy is a genuine paradox. AI is accelerating electricity demand growth. At the same time, modern grids increasingly depend on AI to operate efficiently, integrate renewables and manage volatility. How APAC resolves this tension will shape both its digital competitiveness and its energy transition.
The International Energy Agency (IEA) projects that global electricity demand from data centers will nearly double, from 485 TWh in 2025 to 950 TWh by 2030, with AI-optimized facilities growing fastest. While that represents only about 3% of global electricity demand, the real challenge is not aggregate scale but local concentration.
Data centers cluster in specific cities and industrial corridors, absorbing a disproportionate share of new power capacity and converting abstract constraints into tangible ones. These constraints include grid connection delays, transformer shortages and limited access to clean power.
AI moves faster than the grid
This physical reality is redrawing the competitive map. Digital leadership is no longer defined solely by algorithms, talent or capital. It increasingly depends on the ability to synchronize digital ambition with physical infrastructure. The IEA highlights a stark asymmetry: AI infrastructure can be deployed in one to two years, while grid upgrades typically require five to ten years and generation assets take even longer.
In APAC, this mismatch is acute and uneven. China is scaling AI as a core pillar of industrial policy. India is positioning itself as a global AI hub. Malaysia is emerging as a major Southeast Asian data center node, while Singapore has cautiously reopened capacity under strict sustainability conditions.
Yet the region is far from uniform: some markets have strong grids but severe land constraints. Others have abundant renewables but insufficient transmission. Many economies still rely on state-backed utility monopolies, slow to adopt flexible tariffs or open-access procurement. Competitive advantage will increasingly favor jurisdictions that modernize regulation and coordinate utilities, hyperscalers and communities around synchronized infrastructure planning.
The missing intelligence layer
AI is also the most powerful tool available to optimize energy systems themselves. The IEA identifies mature use cases such as predictive grid maintenance, renewable generation forecasting and battery optimization. It estimates that AI-enabled efficiency gains might deliver up to 13 exajoules in global energy savings by 2035.
For APAC, this potential is critical. The region faces rising demand, rapid electrification and acute political sensitivity around energy costs and is set to drive roughly 60% of all electricity demand growth through 2050. AI’s most transformative energy impact will occur not inside data centers, but across grids, factories and control rooms.
But progress is uneven. East Asia accounts for most smart meters deployed across the region, while much of Southeast Asia remains in early stages.
Momentum is building. Singapore is targeting nationwide advanced metering by 2026. Australia is accelerating the rollout across its National Electricity Market by 2030. India has approved more than 200 million smart meters under its Revamped Distribution Sector Scheme.
But smart meters are sensors, not solutions. A grid that cannot see that demand in near real time cannot manage flexibility, absorb distributed energy resources or dynamically respond to AI-driven load growth. What APAC still lacks is a comprehensive digital energy stack, paired with the skills and governance to operationalize it. Without this gap, the region risks deploying AI infrastructure faster than the energy intelligence needed to sustain it.
A new agenda for policymakers and energy enterprises
For policymakers, data centers can no longer be treated as passive loads. With the right incentives, they become active grid participants. Large-scale battery systems within data centers are projected to reach 20–25 gigawatts globally by 2030 and can support peak shaving and grid stabilization. Non-urgent AI training workloads can be shifted across time zones or geographies to ease pressure on constrained systems.
For enterprises, responsible AI scaling demands a shift from compute-centric to energy-aware decision-making. Chief information officers (CIOs), chief AI officers and chief sustainability officers must collaborate on three fronts: • Evaluating cloud and colocation partners based on local grid conditions and clean power commitments, not cost alone • Designing software architectures that shift energy-intensive workloads to off-peak periods or lower-carbon regions • Establishing governance metrics that weigh AI business value against infrastructure demand and carbon exposure The lesson from enterprise transformations, including IBM’s own Client Zero program, is that AI value at scale comes from redesigning workflows and modernizing foundations, not deploying isolated use cases. That discipline must now extend to energy.
The AI-energy paradox is ultimately a call for greater intentionality. APAC has an opportunity to pioneer an integrated model where AI infrastructure scales with energy realism and AI is aggressively deployed to make energy systems more adaptive and resilient. The defining question is no longer whether the region can build more AI, but whether it can build the energy intelligence to make that growth sustainable.
About the author: Arun Biswas is the Asia-Pacific Leader for Strategic Engagements at IBM Consulting. He works with C-suite leaders to drive AI-powered transformation across regulated industries—energy, banking, and public services. In the energy sector, his focus is on how AI can accelerate the energy transition: helping utilities modernize operations, improve asset performance, strengthen grid resilience, and integrate renewable energy at scale.