The Cooling Decision Is No Longer an Engineering Detail. It is a development decision. And it needs to be made before the building is designed.

Most of this industry has spent decades treating every data centre as its own problem. A new site, a new client, a new set of drawings built from scratch. That approach is ending, and it is ending quickly.

The scale of demand has changed what a single project even is. Buildings that once delivered 4 to 12 MW are now described as data halls – one component inside campuses measured in the hundreds of megawatts, delivered in phases over years. A developer is no longer commissioning a building. They are commissioning a platform.

• 4–12 MW what a single data centre building once delivered, in full

• 100s of MW scale of today's multi-building, multi-phase campuses

  
  

Elegance versus consistency

A one-off design optimised for a single site is judged on its elegance. A repeatable system is judged on its consistency – whether the third campus performs exactly like the first, whether a vendor substitution in one region holds the same tolerances as another, whether the design can be validated once and trusted forty times over. That is a different discipline, not a harder version of the old one.

Capital is driving this directly. Investors think in portfolios, not buildings – and they reward platforms, not one-off risk.

Sovereign wealth funds, pension funds, and infrastructure investors now sit alongside hyperscalers at the table. They expect predictable cost curves and compressed schedules. The centre of gravity in engineering itself is moving – away from drawings and construction packages, toward supply-chain integration, modularisation, and factory-led manufacturing. Parametric design and configuration engines are no longer efficiency tools. They are becoming the baseline.

What this asks of the engineer

None of this removes the need for engineering judgment. If anything, it raises the stakes on it. A design has to be right the first time, because it will be repeated forty times without the chance to quietly improve it on attempt two. The exacting standard our craft has always demanded now has to be built into the system itself, not just into any single building.

We have always believed that mastering a craft earns the right to a certain confidence – not the kind that oversells, but the kind that shows up in the work. That confidence matters more, not less, when one decision now echoes across a portfolio of facilities rather than a single building.

Be assured.