Southeast Asian countries are reassessing the rapid expansion of data centres driven by the surge in demand for artificial intelligence (AI) infrastructure, as the industry increasingly strains local resources and regulatory frameworks. Governments that once actively courted data centre investments for potential economic benefits are now pausing or tightening oversight amid concerns about grid capacity, land use, and regulatory preparedness.
Indonesia’s West Java province recently ordered a halt to construction on a large-scale AI data centre campus operated by Singapore-based BDX Data Centres. The suspension was prompted by a review revealing outstanding permit issues, just weeks after the project broke ground in late September. In neighboring Thailand, authorities paused progress on 166 data centre projects—including 49 already under construction and 117 awaiting approval—as they work to establish comprehensive standards for resource use, site selection, safety, and economic impact.
Analysts say the rush to attract data centre investment was driven by governments’ desire to secure a foothold in the AI economy and avoid falling behind regional competitors. However, the enthusiasm for attracting these facilities has often outpaced consideration of downstream effects. “Not having enough capacity looked worse than having too much,” said Alvin Camba, a political economist, highlighting how regulatory and infrastructural challenges emerged after the initial investment push.
The rapid buildout is putting significant pressure on local power grids, water supplies, and land availability—resources that attracted investors initially due to their relative abundance and low cost. Malaysia’s Johor state illustrates this shift: once coveted for its inexpensive inputs, it is now becoming more selective in allowing new data centre projects as the existing pipeline burdens electricity and water resources.
Experts emphasize that stricter regulations need not diminish the region’s appeal to investors. Instead, companies prioritize predictable environments where they can secure essential utilities and permits consistently. Singapore’s experience exemplifies this dynamic, with its well-coordinated and reliable infrastructure proving more attractive than more permissive but less predictable settings.
The evolving regulatory landscape may also empower governments to negotiate better terms, including local procurement requirements, renewable energy commitments, technology transfer, or increased investment in domestic AI capabilities. As James Guild, a political economy professor in Indonesia, noted, pauses or slowdowns in project approvals provide regulators time to adapt and potentially extract greater value from incoming investments.
Public opposition to data centre development, while not the primary driver of recent government actions, is also emerging as a factor in project feasibility. In Malaysia, plans for a site near Kuala Lumpur were withdrawn after nearly 300 objections were filed with local authorities, underscoring the importance of addressing community concerns to minimize political risks.
Overall, Southeast Asian governments face the challenge of balancing the economic promise of AI-related infrastructure with the sustainability of their resources and regulatory systems, suggesting a more measured approach may define the next phase of the region’s data centre growth.