Power usage effectiveness What is PUE?
PUE, or power usage effectiveness, is a metric used to measure how efficiently a data center uses energy. It compares the total energy consumed by a data center facility with the energy used by IT equipment such as servers, storage, networking, and compute systems.
A lower PUE generally indicates that more energy is being used directly by IT equipment and less is being consumed by supporting systems such as cooling, lighting, and power distribution.
Time to read: 8 minutes 30 seconds | Published: June 11, 2026
目次
PUE main takeaways
- PUE measures data center energy efficiency by comparing total facility energy to IT equipment energy.
- A lower PUE is generally better because it means less energy is being used by non-IT systems.
- PUE is especially important for HPC, AI, and supercomputing environments where power density, cooling, and IT sustainability are major infrastructure considerations.
What does PUE mean?
PUE stands for power usage effectiveness. It is commonly used by data center operators to understand how efficiently energy is being used across a facility.
PUE can be thought of as a "miles per gallon" metric for the data center because it provides a simple way to evaluate energy efficiency.
How is PUE calculated?
PUE = total facility energy divided by IT equipment energy
Total facility energy includes all energy consumed by the data center, including IT equipment, cooling, lighting, and power distribution. IT equipment energy includes the energy used by servers, storage, networking, and other compute infrastructure.
A PUE of 1.0 would represent ideal efficiency, where all energy is used directly by IT equipment. In real-world data centers, PUE is typically above 1.0 because supporting infrastructure is required to keep systems operating reliably.
Why is PUE important?
PUE is important because data center energy efficiency affects operating costs, sustainability goals, and infrastructure planning. As organizations run more compute-intensive workloads such as AI, HPC, and supercomputing, they need to understand how much energy is used to power IT systems compared with the surrounding facility.
PUE can help organizations evaluate cooling efficiency, identify infrastructure inefficiencies, support sustainability initiatives, and plan for higher-density workloads.
What is a good PUE?
A good PUE depends on the data center design, location, workload type, cooling approach, and facility age. In general, the closer the PUE is to 1.0, the more efficiently the data center is using energy.
However, PUE should not be evaluated alone. Organizations may also consider workload performance, reliability, water usage, carbon impact, and infrastructure utilization.
What affects PUE in a data center?
PUE can be affected by both facility-level and IT-level decisions. Cooling design, power distribution, airflow management, equipment utilization, and workload density can all influence how efficiently energy is used.
In high-density environments such as HPC and AI infrastructure, advanced cooling strategies may play a larger role in improving overall efficiency.
Key factors that can affect PUE include:
- Cooling systems
- Power distribution
- Server utilization
- Facility design and efficiency
- Rack density
- Workload type
- Climate and location
- Airflow management
- Liquid cooling use
- Modular data center design
- Heat reuse
How can organizations improve PUE?
Organizations can improve PUE by optimizing the systems and processes that support IT equipment. This may include improving cooling efficiency, reducing power distribution losses, increasing server utilization, modernizing aging infrastructure, and monitoring energy consumption across the data center.
For high-density workloads, liquid cooling can also help reduce reliance on traditional air cooling and support more efficient thermal management. Using heat from direct liquid systems will reduce PUE and can also support greenhouses, district heating, and industrial processes.
PUE vs. TUE vs. ITUE vs. WUE
| Metric | Full term | What it measures | How to think about it |
|---|---|---|---|
| PUE | Power usage effectiveness | The ratio of total data center facility energy to IT equipment energy | A facility-level efficiency metric that shows how much total energy is required to support IT workloads |
| TUE | Total usage effectiveness | A broader measure of energy efficiency across the full data center environment | A more complete efficiency metric that accounts for both facility energy and IT equipment energy use |
| ITUE | IT usage effectiveness | Energy efficiency within the IT equipment layer | An IT-level efficiency metric that helps evaluate how efficiently compute, storage, networking, and other IT systems use energy |
| WUE | Water usage effectiveness | The amount of water used by a data center relative to IT equipment energy use | A sustainability metric that helps evaluate how much water is consumed to support cooling and data center operations |
How does PUE relate to AI, HPC, and supercomputing?
PUE is especially relevant for AI, HPC and supercomputing environments because they s often require high power density, advanced cooling, and careful facility planning. As organizations scale compute-intensive workloads, data center efficiency becomes an important factor in performance, cost management, and sustainability.
For AI, HPC, and supercomputing workloads, PUE can help organizations understand how infrastructure design, cooling, and energy use affect the overall efficiency of the data center.
How HPE supports data center efficiency
HPE helps organizations modernize data center environments for demanding workloads such as HPC, AI, and supercomputing. With services and solutions across data center modernization, liquid cooling, IT sustainability, high-performance computing, heat reuse, and modular infrastructure, HPE can help businesses design environments that support performance, efficiency, and scale.
PUE FAQs
Can PUE be used for a server room?
Yes. PUE can be used to evaluate the energy efficiency of a server room, data center, or similar IT environment, as long as total facility energy and IT equipment energy can be measured.
For smaller server rooms, PUE may be harder to calculate precisely because cooling, lighting, and power usage may be shared with the broader building.
How do data centers monitor PUE?
Data centers monitor PUE by measuring total facility energy and IT equipment energy over time. This can be done through metering, power monitoring systems, environmental sensors, DCIM software, or manual calculations using energy consumption data.
Continuous monitoring provides a more accurate view than a one-time calculation.
Why is my data center PUE high?
A high PUE may be caused by inefficient cooling, poor airflow management, low server utilization, older infrastructure, power distribution losses, or facility design limitations.
High-density workloads such as AI and HPC can also increase cooling demands if the data center is not designed for them.
Does liquid cooling improve PUE?
Liquid cooling can help improve PUE in high-density environments by reducing reliance on traditional air cooling. Its impact depends on the workload, facility design, rack density, and cooling architecture.
Liquid cooling is especially relevant for AI, HPC, and supercomputing environments.
When should an organization consider a PUE assessment?
Organizations may consider a PUE assessment when energy costs are rising, cooling capacity is constrained, sustainability goals are difficult to measure, or new high-density workloads are being planned.
A PUE assessment can help identify inefficiencies in cooling, power distribution, and infrastructure design.