What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
European data centres can reduce some of the energy used to cool servers, shift when they draw power, and put waste heat to work. The practical options include cooling-optimization software, liquid cooling, battery dispatch and heat recovery—but they do not all reduce a facility’s total electricity consumption, and their reported results cannot be treated as a single Europe-wide savings figure.
Why data-centre energy demand is under pressure
Data-centre electricity use in Europe is projected to rise from 96 terawatt-hours (TWh) in 2024 to 236 TWh by 2035, according to Ember, as reported by Euronews in 2026. That projection covers European data centres overall; it is not an AI-only forecast. It helps explain why operators are looking for efficiency and flexibility measures, but it does not show how much any one intervention can save.
As an Amazon Associate I earn from qualifying purchases.
AI workloads add to the pressure because they rely on energy-intensive computing equipment and the infrastructure needed to keep it operating. The examples below address different parts of that infrastructure. Some report less energy for cooling, some lower costs, some change the timing of grid demand, and some recover heat. Those outcomes are not interchangeable.
What the European projects are doing
| Approach and example | Reported result | What the result measures |
|---|---|---|
| Cooling sensors and digital twins: EkkoSense at Virgin Media O2 sites | Average 15% cooling-energy reduction; more than £1 million in annual savings and 760 tonnes of location-based CO2 equivalent, as reported by Euronews in 2024 | Operator-reported outcomes across 20 sites, not a universal performance promise |
| Cooling sensors and digital twins: Telefónica Germany project using EkkoSense | Initial evaluations estimated a 15–20% reduction in cooling-system energy consumption | An initial estimate for this program |
| Cooling-control software: etalytics at NTT’s Bonn data centre | Chiller electricity use fell 19.1% in the first months of the trial; NTT expected savings of up to 25% over a full year | A reported trial result and a separate full-year expectation |
| Cooling-control software: OctaiPipe ACE at an Italian data centre | Cooling energy costs fell by up to 30%, and total site infrastructure energy efficiency improved by more than 10% | Figures in Italtel’s deployment announcement for a 2 MW carrier-neutral facility |
| Immersion cooling: Submer systems, as described by Euronews | Telefónica claimed energy-efficiency improvements of up to 50% | A customer claim; “up to” is not a general result for all facilities |
| Battery dispatch: GridBeyond at two Dublin data centres | Combined 8 MW of flexible capacity | Grid flexibility, not a measured reduction in total electricity consumption |
| Waste-heat use: Deep Green swimming-pool trial in southwest England | The pool was expected to cut gas use by 62%, save more than £20,000 annually and reduce annual emissions by 25.8 tonnes | Expected benefits for the heat user, not lower data-centre electricity use |
| Hybrid cooling and heat recovery: HEATWISE project pilots | Modelled cooling-load reductions of up to 53.2% and heat-recovery improvements of up to 17.1% for specific scenarios | Simulation or project findings tied to particular pilot conditions |
| Facility cooling: Scaleway’s Paris DC5, in an NVIDIA case study | Electricity consumption reduced by 30–40% and water use by 9–10 times versus traditional data centres | A vendor case-study claim about the facility, not necessarily an isolated AI workload |
Cooling controls can improve how existing equipment is used
Sensors and a digital twin reveal hot and cold spots
Cooling systems are often operated to protect the hottest or most vulnerable equipment, which can leave other areas overcooled. EkkoSense uses wireless sensors to track room temperature, airflow, power and cooling capacity. Its software represents those readings in a three-dimensional view of the facility, helping operators identify over- and under-cooled equipment and adjust the system.
At Virgin Media O2, Euronews reported an average 15% reduction in cooling energy across 20 sites in 2024, alongside the operator-reported annual savings and emissions figures shown above. That result belongs to the sites and reporting context described; it should not be assumed for a different data centre. In a separate Telefónica Germany project, the company described EkkoSense’s use of IoT sensors, analytics and a real-time 3D digital twin. Telefónica said initial evaluations estimated 15–20% lower cooling-system energy use. The system is also intended to map thermal and load risks and recommend changes using existing site capacity.
Software can optimize chillers, pumps and heat exchangers
German startup etalytics models cooling equipment such as chillers, pumps and heat exchangers, then calculates operating options that remain within temperature limits. Euronews reported that NTT’s Bonn data-centre trial reduced chiller electricity use by 19.1% in its first months. NTT expected savings of up to 25% over a full year; that figure is an expectation, not a completed full-year result.
OctaiPipe’s AI for Cooling Efficiency (ACE) software uses a digital twin and collaborative AI agents to monitor and optimize cooling parameters. Italtel says ACE integrates with existing building management systems (BMS) and data-centre infrastructure management (DCIM) systems without requiring new sensors. Its deployment announcement describes a 2 MW carrier-neutral Italian data centre where cooling energy costs fell by up to 30% and total site infrastructure energy efficiency improved by more than 10%. These are deployment-announcement figures, not independently audited results that establish what another operator will achieve.
Liquid and immersion cooling remove heat closer to the servers
Air cooling moves heat away using fans and air-conditioning equipment. Immersion cooling instead places servers in a non-conductive liquid that carries heat away directly. Submer’s systems use this approach; Euronews quoted Telefónica as claiming energy-efficiency improvements of up to 50%, as well as reduced refrigerant-related carbon emissions, footprint and total cost of ownership. The “up to” figure is the customer’s claim, not a guaranteed saving or a result that can be applied to every data centre.
Liquid cooling can also be designed to make server heat available for reuse. The Bitten AI supercomputer at the University of Southern Denmark, developed with Danfoss and HPE, uses liquid cooling with full heat recovery and connects recovered heat to the local energy system. It demonstrates one project configuration; the case study does not establish a general electricity-savings percentage.
These are facility-scale engineering approaches, not consumer PC cooling accessories. Their suitability depends on the site’s existing design, operating requirements and ability to integrate new cooling and heat-transfer equipment.
Rank #3
Batteries can shift demand without reducing total electricity use
Data centres commonly keep batteries for backup power. GridBeyond’s software coordinates battery charging and discharging in response to grid conditions. Euronews reported that it was installed at two Keppel DC REIT data centres in Dublin, providing a combined 8 MW of flexible capacity.
That capacity describes the ability to adjust when electricity is drawn from the grid. Shifting demand may help a facility respond to grid conditions, but it does not by itself show that the facility consumed fewer kilowatt-hours overall. The 8 MW figure should therefore not be compared with a percentage reduction in cooling energy or total site use.
Recovered heat can benefit nearby users
Heat reuse depends on a nearby, suitable customer
Server heat can be transferred to buildings or energy systems that need heat. UK startup Deep Green places computing facilities near potential users, such as swimming pools or district-heating networks, and transfers server heat using warmed liquid. In a southwest England swimming-pool trial reported by Euronews, the pool was expected to reduce gas use by 62%, save more than £20,000 a year and cut annual emissions by 25.8 tonnes. Those are expected benefits for the pool; they are not reductions in the data centre’s electricity consumption.
For heat recovery to be useful, there must be a nearby user with a suitable demand profile and a workable connection. A facility’s ability to produce heat does not guarantee that a customer can use it when it is available.
European pilot modelling explores cooling and heat recovery together
The European Commission’s HEATWISE periodic report covers pilots in Denmark, Poland, Switzerland, Sweden and Türkiye. The work includes hybrid cooling, digital-twin data management, building energy management and heat recovery. It reports modelled cooling-load reductions of up to 53.2% and heat-recovery improvements of up to 17.1% for specific scenarios. These are simulations or project findings tied to particular pilot conditions, not observed results that can be generalized to all European sites.
Free tools Windows power users keep installed
One-click scans. No signup required.
How to judge an energy-saving claim
Before comparing two projects, check whether they measure the same thing and use the same boundary. A lower cooling bill, less chiller electricity, a reduction in whole-site consumption and a grid-flexibility commitment answer different questions.
- Outcome: Is the figure about cooling energy, total site electricity, operating cost, peak grid draw, water use or recovered heat?
- Boundary: Does it cover the cooling plant, a data-centre room, the whole facility or a nearby heat network?
- Evidence status: Is the result measured in a deployment, an initial estimate, a customer or vendor claim, a simulation or a forecast?
- Operating conditions: What are the facility size, climate, workload, cooling design and baseline? Results can vary with these conditions and with how well the site was optimized before the change.
- Implementation: Does the approach use existing controls, need additional sensors or hardware, require integration with BMS or DCIM, or affect uptime planning?
For example, OctaiPipe’s reported reduction in cooling energy costs, GridBeyond’s flexible capacity in megawatts and HEATWISE’s modelled cooling-load result are not comparable savings measures. Likewise, a claim about a facility’s water use or a swimming pool’s gas use does not establish a reduction in AI computing electricity.
What these examples establish—and what they do not
European projects show several credible intervention points: better control of existing cooling systems, liquid-based heat removal, battery-enabled demand flexibility and reuse of server heat. Some deployments report lower cooling energy or costs; other examples describe an expected benefit, a modelled result or a capability to shift grid demand.
The available examples do not establish one verified percentage reduction in total AI data-centre electricity demand across Europe. The Europe-wide projection cited here covers all data centres, while the project figures use different measures, boundaries and evidence types. A meaningful comparison needs those distinctions alongside the number.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
