From Commitments to Capacity: Bottlenecks, Mitigation Pathways and Beneficiaries

The key debate around AI and data center infrastructure is shifting from capacity commitments to capacity conversion. Recent community pushback in the US demonstrates that demand alone is no longer sufficient to guarantee data center growth, with power affordability, water consumption, emissions, noise and permitting increasingly influencing project outcomes. As operators scale multi-gigawatt AI campuses globally, we see debates increasingly emerging across APAC. We identify 10 capacity conversion bottlenecks among three categories: Hard Infrastructure Bottlenecks (grid connection & reliability, power equipment supply, optical networking, land availability — risk elevated in Hong Kong, Singapore, Tokyo); Emerging Constraints (ambient environment, water demand, skilled labor, infrastructure-ready land — risk elevated in Indonesia, Thailand, Malaysia);