How Much Electricity and Water Does an AI Data Centre Use? Racks, PUE, WUE, Cooling and Heat Reuse Explorer

What Are PUE and WUE?

Power Usage Effectiveness (PUE) is a metric used to measure the energy efficiency of a data centre. It represents the ratio of total power entering the facility to the power used directly by IT equipment. A PUE of 1.0 would indicate that all the power consumed goes directly into running the computers, while higher values represent inefficiencies in the cooling and other auxiliary systems.

Water Usage Effectiveness (WUE) is a similar metric but for water consumption, focusing on how much water is used per unit of IT load. It helps assess the efficiency of water usage in data centres, especially important given the increasing focus on sustainability.

Impact of Cooling Methods

The choice between air and liquid cooling significantly impacts both PUE and WUE. Air cooling is simpler but less efficient, often leading to higher energy consumption due to the need for more powerful fans and larger HVAC systems. Liquid cooling, on the other hand, can be much more efficient as it directly transfers heat away from the processors, reducing the load on air conditioning units.

Direct-to-chip liquid cooling, in particular, is highly effective because it cools individual components at a microscopic level, leading to lower temperatures and higher efficiency. However, it requires specialized infrastructure and can be more expensive.

Energy and Water Consumption

The amount of electricity consumed by an AI data centre is directly related to the number of GPU racks and their power consumption per rack. For instance, a single 130 kW rack can consume significantly more energy than a 10 kW rack. This difference in power usage can greatly affect PUE values, as higher power consumption often correlates with less efficient cooling systems.

Water usage is also influenced by the cooling method chosen. Air cooling typically requires more water for evaporative cooling processes, while liquid cooling can reduce overall water use but may still require significant amounts depending on the system's design and efficiency.

Heat Reuse and Sustainability

Efficient data centres are increasingly exploring ways to reuse the heat generated by their servers. This not only reduces energy consumption but also has environmental benefits. For example, in some cases, the warm water from liquid cooling systems can be used for heating buildings or greenhouses, turning a potential waste product into a valuable resource.

However, the feasibility of heat reuse depends on local infrastructure and regulations. Data centres located near industrial facilities might find it easier to implement such systems compared to those in more remote locations.

Frequently asked questions

Try it live

Everything above runs in your browser — open How Much Electricity and Water Does an AI Data Centre Use? Racks, PUE, WUE, Cooling and Heat Reuse Explorer and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.