DATA CENTRE COOLING
Do data centres make good neighbours?
Data centres are the essential infrastructure of our age. Due to their power (and sometimes water) hungry nature, however, exacerbated in recent years due to the extra power and cooling requirements of AI servers, data centre planning and construction must come with a focus on reducing heat rejection and reusing heat, the by-product of cooling demand. Tim Mitchell, Sales Director for Klima-Therm, argues that this heat should be viewed as a municipal resource in order to off set at least some of the data centres’ excesses.
"For QMUL, this has led to a signifi cantly reduced
dependence on
conventional gas boilers."
D
ata centres already consume very large amounts of electricity to run and to keep themselves cool, and the rise of AI is dramatically increasing this consumption.
The IEA and energy analysts have identifi ed that a wave of new AI infrastructure across the world is driving electricity demand at an alarming rate. Signifi cantly more energy-intensive than a standard Google search, generative AI uses 10 to 33 times more electricity per query, and the latest versions, such as GPT-4, use 50 times more electricity than earlier iterations – a growing concern for electricity grids. While it’s clear and largely unavoidable that data centres
are ‘consumers’, what’s wasted from this process – heat – can be used by adjacent buildings, potentially creating a symbiotic relationship that reduces the carbon footprint of the heat users and actually improves data centre effi ciency.
Living on the edge To get the most out of data centres, we need to build them in
locations where heat is needed. These aren’t the huge, ‘hyper- scale’ operations more commonly found in America, more the smaller, edge solutions which can sit alongside urban and industrial developments. For district heating systems to work at their optimum, the pipe delivering the heat needs to be as short as possible to prevent heat losses. Thermal storage and smart controls are also essential, as while data centres run at the same pace all year round, heat needs ebbs and fl ows with the changing seasons. When a data centre supplies a nearby heat network, the
operator can reduce on-site energy withdrawal for cooling and potentially monetise the heat. For the heat network, a reliable, high-grade source lowers the carbon intensity of heating and can reduce the need for gas or expensive electrifi cation at times of high demand. From a system perspective, bringing heat recovery into project appraisals changes the investment case for where and how to build new data centres in cities. Despite the obvious benefi ts, there are still policy and
regulatory hurdles that need to be addressed if heat recovery is to reach its potential. Planning policies should encourage co-locating centres with heat demand, procurement and tariff structures, allowing data centre operators to be paid for the heat they export. Water regulation should also recognise the trade-off s between cooling water use and reuse opportunities. Recent Government funding and pilots in the UK show there is an appetite to make these policy changes, but scaling will require clearer incentives and a recognition that data centres can be energy assets, not just consumers. Proven examples where this approach is working well include Queen Mary University of London (QMUL). Waste heat from the onsite data centre is used to create hot water, reaching temperatures as high as 75°C, which is directly integrated into the existing plumbing system, servicing taps
12 November 2025 •
www.acr-news.com Download the ACR News app today
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40
