Turning data centre waste heat into an investment-grade district export scheme

The situation

A leading European data centre operator was being asked the same question by three different audiences. Planners wanted to know what would happen to the heat its new capacity would reject. The local authority wanted district heating options examined. And its own sustainability function needed an answer for the waste heat utilisation and energy reuse factor lines that EU data centre reporting now requires.

The honest answer was that nobody knew. Heat reuse had appeared in planning statements as an aspiration, but the site had never been mapped, no offtaker had been engaged, and earlier desk studies were too shallow to take to an investment committee. With grid capacity tightening across the region and the first EED submissions already filed, the question was no longer going to wait.

The constraint

Data centre heat is plentiful but humble. With an air-cooled architecture, most of it leaves at temperatures too low to feed a heating network directly, which means a heat pump lift, which means capital cost and an electricity bill the scheme’s economics have to carry. Distance matters too: pipework to a remote offtaker can sink a scheme before the first kilowatt hour moves.

Demand profiles cut the other way. A heating network wants most of its heat in winter; a data centre rejects heat at a near constant rate all year. Any credible scheme had to be honest about what happens to the summer surplus and what that does to the economics.

The operator also needed the analysis to survive contact with three very different counterparties: a municipal heating development, a neighbouring campus and a private industrial user. A generic feasibility gloss would persuade none of them.

What EM3 engineered

EM3 began at the source. Rejected heat was mapped across the site: quantities, temperature levels and availability profiles across the year, built from metered cooling plant data rather than nameplate assumptions. Data centre heat has one structural advantage over almost every other industrial source, and the mapping quantified it: it is available 365 days a year, with a profile that barely moves.

The study then assessed liquid cooling readiness, because the cooling architecture decides the export case. Direct-to-chip and rear-door heat exchanger options for the operator’s higher density racks would raise return water temperatures, improving chiller efficiency on site and reducing the heat pump lift the export scheme needs.

Three offtake routes were developed in parallel: the municipal district heating development, the neighbouring campus and the industrial user. For each, EM3 sized the heat pump plant and pipework, modelled capital and operating costs with sensitivity ranges, structured the commercial options, and quantified the carbon case from the fossil heat the export would displace. A risk register and delivery phasing completed the package.

The results

The operator now holds an investment-grade feasibility study: one preferred scheme, fully costed with sensitivities, two credible fallbacks, and a quantified carbon case it can put in front of planners, offtakers and its own board with equal confidence.

Investment-grade meant exactly that: itemised capital and operating costs for each route, sensitivity analysis on power prices and heat tariffs, a phased delivery plan with named decision points, and a risk register the operator’s engineers had challenged line by line before it went anywhere near the board.

The reporting question answered itself along the way. The site’s energy reuse factor now has a measured baseline and a costed path upward, which turns a line in the EED submission from an awkward zero into a documented plan.

What it means for the sector

Waste heat has moved from public relations to formal KPI. Under Delegated Regulation 2024/1364, energy reuse is reported, visible and comparable, and planners are reading the numbers.

The operators who map their heat early will shape the district heating schemes around them. The ones who wait will be retrofitting answers to questions other people have already framed. Heat mapping is cheap, fast, and the difference between the two positions.