Data Centre Insights – May 2026
Early enterprise adoption of AI is reshaping how data centre performance is measured. Traditional metrics such as megawatts and cost per kilowatt remain important, but are now accompanied by emerging measures, with attention shifting towards tokens - the fundamental units of AI output. This marks a move from capacity-based to outcome-driven measurement.
In this context, AI data centres are increasingly viewed as “AI factories”, converting energy into intelligence. Efficiency is no longer defined solely by power consumption, but by how effectively that power translates into usable AI output and value.
This shift is also influencing cost dynamics, with cost per token emerging as a more precise indicator. However, variability across workloads and infrastructure means traditional metrics remain relevant, with both approaches used in tandem to provide a more complete view of performance and economics.
Download the full Data Centre Insights External Link to explore how evolving metrics - from megawatts to tokens - are reshaping AI infrastructure economics and what this means for operators and enterprises navigating the next phase of AI adoption.
DC Insights is Cushman & Wakefield’s monthly market brief from the Asia Pacific Data Centre Group, offering timely insights on market movements, key developments and outlooks across the region.
