AI Data Centre Electricity Calculator
Estimate the annual electricity consumption, energy costs and carbon impact of an AI data centre based on server count, utilisation and operating assumptions. An educational planning tool for businesses, researchers, journalists and policy analysts.
Important: These tools provide educational estimates only. Actual AI electricity use varies by model, data centre, cooling system, hardware, location and workload. Do not use these results as official energy, planning, investment or engineering advice.
Facility assumptions
What this consumption is equivalent to
Insights
- This facility would consume more electricity than approximately 2,384.7 UK homes use in a year.
- Annual electricity costs could reach approximately £1,609,650.00 — over £1.6 million.
- Improving PUE to 1.2 (Excellent) could save approximately £229,950.00 annually in electricity costs.
- At UK grid average carbon intensity, this facility would emit roughly 1,332.8 tonnes of CO₂ per year.
- What does this mean for the UK grid? Large AI facilities place sustained, round-the-clock demand on local networks, which can require grid reinforcement and influence where new capacity is built.
What is a data centre?
A data centre is a facility housing large numbers of servers that store data and run software. AI data centres are specialised for training and running machine-learning models, packing in power-hungry GPU and accelerator hardware that runs continuously.
Why AI consumes large amounts of electricity
AI workloads rely on dense clusters of high-performance processors that draw far more power than traditional servers. Training and serving large models keeps that hardware busy around the clock, and every watt of compute also generates heat that must be removed — adding further energy demand.
What is PUE?
Power Usage Effectiveness (PUE) compares the total energy a facility uses against the energy delivered to its IT equipment. A PUE of 1.0 would be perfectly efficient; 1.2 is excellent, 1.4 is good, 1.6 is average and 2.0 is poor. The higher the PUE, the more energy is spent on cooling and overheads rather than computing.
How cooling systems affect energy use
Cooling can account for a large share of a data centre's overhead. Air cooling, evaporative cooling, liquid cooling and free-air cooling all carry different energy and water trade-offs. Efficient cooling and a cooler climate reduce PUE and overall electricity demand.
Why AI infrastructure matters to UK energy policy
As AI adoption grows, data centres are becoming a meaningful and concentrated source of electricity demand. This affects grid planning, where new capacity is built, electricity pricing and the UK's net-zero commitments — making AI infrastructure an increasingly important part of national energy policy.
How This Calculator Works
IT load (kW) = servers × watts per server × utilisation ÷ 1,000. Facility load = IT load × PUE. We multiply facility load by 24 hours to get daily kWh, then scale to monthly (×30) and annual (×365). Costs use the electricity price you enter. Carbon is estimated using the UK grid average of 0.207 kg CO₂ per kWh.
Equivalents assume a typical UK home uses 2,700 kWh per year, a full EV charge is about 54 kWh and a kettle boil is around 0.1 kWh.