News and Insights

At DDI 2026 — the Danish Data Center Industry conference — a panel discussion tackled a question that’s becoming harder to ignore: what actually happens to the heat a data centre produces?

The instinctive answer is that surplus heat is a resource — capture it, feed it into a district heating network, and the energy math improves. In practice, that’s only true under specific conditions. Surplus heat is not automatically valuable.

What Determines Whether Heat Is Usable

Heat leaving a data centre is only useful to a district heating network if it meets that network on its own terms. Four factors decide whether it does:

• Temperature — is the heat hot enough to be useful without further upgrading?
• Flow — is there enough volume to matter at a network scale?
• Distance — how far does the heat have to travel before it reaches a facility or community that can use it?
• System design — is the surrounding infrastructure actually built to receive and integrate it?

Miss any one of these, and “waste heat recovery” becomes a slide in a sustainability report rather than something a utility can actually put to use. This is why cooling architecture — not just the presence of surplus heat — is what determines whether reuse is realistic.

Why Return Temperature Changes the Economics

The clearest lever in that list is temperature, and the difference is larger than it might seem on paper.

Take two return temperature profiles: 22°C/32°C versus 22°C/42°C. The second isn’t just “10 degrees better” — it changes the underlying economics. Higher return temperatures mean lower required flow rates for the same heat transfer, which means smaller infrastructure to move that heat. It also means more hours per year where free cooling is viable, and heat that arrives at a district network hot enough to be genuinely useful rather than requiring additional energy to upgrade it further.

None of this is theoretical — it’s the difference between heat that a utility can integrate directly and heat that costs more to use than it’s worth.

Real Performance Happens Under Real Conditions

There’s a second problem that’s easy to overlook: cooling systems are rarely operating at their rated, ideal performance point. Real-world operation happens at partial loads, under changing workloads, with thermal behaviour that doesn’t hold steady.

This is where precision and response time stop being abstract specifications and start mattering directly for heat reuse. A system that can’t hold a stable return temperature under a shifting load doesn’t just risk local cooling performance — it risks the consistency that a district heating partner needs to plan around.

What This Means for Cooling Architecture

Physics doesn’t scale with assumptions. A cooling system designed around a single ideal operating point will behave differently — and less predictably — once it’s handling the partial loads and load changes that make up most of its actual operating life.

Stable thermal behaviour under real, changing conditions isn’t a sustainability add-on — it’s what determines whether heat reuse is a realistic option or a best-case scenario that rarely holds up outside a datasheet.

Data centres and district heating networks don’t automatically make good partners. Cooling design is what decides whether they do.

This piece draws on a panel discussion at DDI 2026 featuring NGC and DIN Forsyning.