As artificial intelligence workloads accelerate, data centre growth is no longer limited by compute or construction speed—it is constrained by power availability. In a recent Data Centre Magazine article, Rehlko Chief Brand and Sustainability Officer Francis Perrin outlines how AI is fundamentally altering energy demand profiles and exposing the limits of traditional infrastructure planning.

AI workloads require sustained, high‑density power in always‑on environments, placing continuous strain on systems originally designed for predictable, intermittent demand. At the same time, utilities are struggling to keep pace. Grid congestion, multi‑year interconnection delays, and regional capacity shortages are slowing projects that would otherwise be ready to deploy. Increasingly, the key question for developers is no longer “can we build it?”—but “can we power it?”.

This shift is forcing a rethink of the role power plays in data centre design. Backup infrastructure can no longer function as passive insurance for rare outages. In AI environments, power systems become active performance enablers—directly impacting uptime, efficiency, and long‑term reliability. Even minor disruptions or fluctuations can carry outsized consequences when margins for error effectively disappear.

As AI continues to outpace the infrastructure designed to support it, power has become a strategic priority—not a downstream consideration. Read the article to lean how he industry’s ability to scale will depend not just on how quickly capacity is deployed, but on how reliably that capacity can be sustained over time. In the age of AI, sustainability is no longer about intent alone—it is about endurance.