The AI Power Crunch: Why Five ASEAN Nations Are Going Nuclear

The AI Power Crunch: Why Five ASEAN Nations Are Going Nuclear

Southeast Asia has never produced a single watt of nuclear energy. That is about to change. Five ASEAN member states, Indonesia, Malaysia, Thailand, Vietnam, and the Philippines, are actively pursuing nuclear power programmes, driven not by traditional energy needs but by the staggering electricity demands of artificial intelligence data centres. A standard AI data centre consumes as much power as 100,000 households, and the region's more than 2,000 facilities are just the beginning.

The Scale of the Problem

The numbers tell a stark story. Singapore, Malaysia, and Indonesia collectively host data centre clusters expected to account for 40% of global data capacity by 2030. Microsoft alone committed $5.5 billion to Singapore this week, with massive infrastructure expansion built into the deal. Google, Nvidia, and dozens of smaller players are pouring billions more into the region.

But AI infrastructure needs power, and lots of it. The region's current energy mix, heavily dependent on natural gas and coal, cannot scale fast enough to meet projected demand. The conflict in the Middle East has made matters worse, with crude oil price surges underscoring the vulnerability of Asia's energy supply chains.

There is a more serious, new and growing momentum for the development of nuclear energy in Southeast Asia.

Country by Country: The Nuclear Sprint

Each ASEAN nation is charting its own nuclear path, shaped by local politics, geography, and economic ambitions.

Vietnam is furthest ahead. The country is building two nuclear plants backed by Russian state corporation Rosatom, described by Prime Minister Pham Minh Chinh as nationally significant, strategic projects. A revised atomic energy law took effect in January 2026, clearing regulatory obstacles. Vietnam is also breaking ground on a $2.1 billion AI data centre campus in Ho Chi Minh City on 30 April, Liberation Day, a deliberate symbolic choice.

Indonesia added nuclear to its national energy plan in 2025, targeting two small modular reactors by 2034. The country is in discussions with Canadian and other international partners to deliver the technology.

Thailand set a target last year of adding 600 megawatts of nuclear generating capacity by 2037, positioning nuclear as a key component of its long-term energy strategy.

The Philippines approved a nuclear roadmap for potential investors in February 2026, with a target of 2032 for its first operational plant.

Malaysia is reviving previously mothballed nuclear plans explicitly because of AI data centre ambitions. With more than 500 operational data centres, 300 under construction, and 1,140 planned, the country's power grid faces unprecedented strain.

The Geopolitical Dimension

Nuclear power in Southeast Asia is not just an energy story. It is a geopolitics story. Vietnam's partnership with Rosatom ties its energy future to Russia. Indonesia's discussions with Canada and other Western partners reflect a different alignment^✦. The choice of nuclear technology partner carries decades-long strategic implications.

Meanwhile, the AI data centre boom is creating new dependencies. Companies building facilities in the region, from Baidu's Xiaodu expanding into Thailand and Singapore to global hyperscalers, need guaranteed power supplies. A nuclear plant takes a decade to build. An AI data centre takes two years. That timing mismatch creates a bridge period where fossil fuels will fill the gap, complicating the region's climate commitments.

• Vietnam leads with active nuclear construction and a revised legal framework
• Indonesia is pursuing small modular reactors as a flexible, lower-risk option
• Thailand's 600 MW target is modest but represents a political breakthrough
• The Philippines' 2032 target is ambitious given the country's limited nuclear experience
• Malaysia's decision to revive nuclear plans reflects the urgency of AI-driven^✦ power demand

The Renewable Alternative

Not everyone is convinced nuclear is the answer. Solar and wind capacity is expanding rapidly across the region, and battery storage technology is improving. Critics argue that the decade-long construction timelines for nuclear plants make them poorly suited to an AI industry that is evolving on 18-month product cycles.

But renewables face their own challenges in Southeast Asia. Intermittency, land constraints, and grid integration issues make them difficult to scale as baseload power for facilities that must run around the clock. Nuclear offers the one thing AI data centres need most: reliable, continuous power.

Southeast Asia's data centre growth is creating power demands that no single energy source can meet. Nuclear is part of the answer, but not the whole answer.