November 26, 2025
Deep Green
Power Usage Effectiveness: The metric that CSOs and CTOs alike lose sleep over. Your colocation solution’s PUE has the potential to save you costs on your high-performance compute and reduce your enterprise’s emissions to boot.
But what is the magic PUE metric that you should be looking for in a colocation partner? How is PUE calculated? And is data centre PUE the only way to measure compute efficiency? Sit back and relax, because today, we’re covering everything you need to know.
What is data centre PUE?
Introduced by The Green Grid in 2007, PUE is designed to measure the efficiency of data centre facilities. PUE is calculated by dividing a data centre’s total energy consumption by the energy that powers its compute:
PUE = Total energy ÷ Compute energy
The compute energy refers to the power that’s used to run servers, networks and storage. The total energy includes the compute energy, plus the power needed for cooling systems, lighting, security, and uninterruptible power supplies.
The result? A figure that functions as a ratio, indicating how efficiently a data centre runs its compute, compared to how much power it uses on things other than compute.
Ideally, your colocation data centre will use as little extra power as possible, resulting in a low PUE. The less power a data centre uses to run your high-performance compute (HPC), the lower your operational costs and carbon emissions. The industry benchmark for PUE sits between 1.225 and 1.7. So, partner with a UK data centre whose PUE is at the lower end of this scale, or even beneath it entirely.
How data centres can reduce PUE
Now you know that a low data centre PUE is what you’re after from your colocation partner. But what affects PUE? And how can data centres keep the metric as low as possible?
Essentially, all of the power needed to run a data centre facility is part of the PUE calculation. Reducing this power will, in turn, reduce PUE.
Start with a data centre’s cooling system. Some facilities use air-cooled, while others opt for liquid cooling. Direct liquid cooling offers many advantages over air cooling: it can draw heat away from hardware more efficiently than air, and requires less energy to power. It also allows servers to be packed in more closely, so more servers can be cooled at once. All of this leads to a lower PUE.
Another factor influencing PUE is the way a data centre manages its airflow. Facilities with purpose-built hot and cold aisle containment (a system that prevents hot and cold air from mixing) improve the efficiency of their cooling.
How a low PUE can impact your enterprise
Why is a data centre with a low PUE the best colocation partner for your enterprise? There are two key benefits to consider. The first is that a lower PUE means lower operating costs for the facility in question. This results in cheaper colocation, so you can scale your HPC while staying on budget.
Then, there are the environmental benefits. Improving data centre efficiency means less energy is needed to run your servers, reducing your compute emissions. Not only does this become part of your enterprise’s ESG narrative, but it also reduces your scope 2 or 3 emissions. Which scope they belong to depends on whether you’ve purchased the energy to run your compute directly from the grid (calculated as scope 2) or whether your colocation partner charges you for the energy (calculated as scope 3).
Want more advice on meeting ESG guidelines and reducing scope 1-4 emissions? Our checklist for smart CSOs is here to help.
What metrics should UK data centres measure beyond PUE?
So, you’ve chosen a colocation partner with a low PUE. Job done, right? Not so fast. Data centre PUE isn’t the be-all and end-all when it comes to efficiency or environmental impact. There are other key metrics that speak to a facility’s sustainability.
First, consider a facility’s Energy Reuse Factor (ERF). This metric shows how much of a data centre’s energy is reused, for example, heat that’s passed onto district heating systems. ERF is a percentage score; a higher score is better, indicating that more energy is being reused. Aside from numbers, look at how a data centre collaborates with the community beyond donating heat for a holistic picture of their community impact.
When it comes to heat reuse, let’s clear something up. There’s a common misconception that passing on heat requires more power, negatively impacting a data centre’s PUE. This isn’t the case: data centres can use passive cooling techniques to help heat pumps reach operational temperatures, rather than increasing energy input. This being said, we’d like to introduce our very own metric that can show a truer picture of a data centre’s efficiency: Net PUE.
Net PUE = (Total facility power – Heat reuse) ÷ Compute power
This metric goes beyond measuring a facility’s efficiency, instead showing the impact the data centre has on energy efficiency in the wider community.
If we only measure data centres on their energy efficiency, it would be pretty simple for facilities to choose more efficient liquid cooling systems, reduce their PUE, and be done with it. But this doesn’t factor in how data centres can negatively impact local communities by draining their water resources. The Water Usage Effectiveness (WUE) metric combats this by showing how much water a facility uses compared to the energy that powers their compute. The lower the WUE, the less water is used: industry standards sit between 0.3 and 0.8. At Deep Green, our water usage and corresponding WUE are practically zero.
How Deep Green achieves industry-leading efficiency
In the words of Deep Green’s Chief Technology Officer Matt Cragg, “We’re aiming to become the most efficient data centre operator in the world.”
This ambition means our facilities go way beyond standard PUE measurements and efficiency improvements. Through our heat reuse programs, we donate heat from our servers to local spaces, homes and businesses via district heating systems.
And it’s not just about our community impact; when it comes to the metrics, we outperform industry standards on all counts:
Metric | Deep Green | Typical data centre |
|---|---|---|
PUE | 1.03-1.07 | 1.225-1.7 |
Density (kW/rack) | 20-480 | 3-12 |
Water Usage Effectiveness (WUE) | 0.00018 | 0.3-0.8 |
Energy Reuse Factor (ERF) | 82-93% | 0% |
Renewable Energy Factor (REF) | 100% | <50% |
Cooling Efficiency Ratio (CER) | 2+ | <1 |
How important is PUE when choosing a colocation partner?
We’ll be honest: you need a low data centre PUE if you want to reduce operational costs and reduce your scope 2 or 3 emissions. But PUE is only the starting point when it comes to sustainability. To reduce scope 4 emissions and build your ESG story, choose a data centre with a strong ERF and net PUE, too. These metrics speak to how local communities benefit from donated heat, indicating holistic efficiency not just in the data centre itself, but beyond into the wider community.
Want to see how our heat reuse projects come to life?




