As AI adoption accelerates across APAC, data centres are under pressure to become “AI-ready” quickly. But AI readiness should not simply mean adding more GPUs, more power, and more cooling. It requires a shift from capacity planning to capability planning: understanding how AI workloads behave, how inference demand fluctuates, how agentic systems create unpredictable compute usage, and how governance can reduce unnecessary infrastructure pressure. This session presents a practical framework for AI-ready data centre governance, covering workload observability, compute efficiency, model-routing discipline, security risk, and cross-team decision-making between infrastructure, platform, and AI teams. Attendees will learn how to distinguish real capacity needs from avoidable workload waste, how to plan for AI growth without overbuilding, and how to align AI infrastructure investments with sustainability, resiliency, and operational control.