What is a Data Center? 

A data center is a building that houses computers and equipment used to store, process, and send information for things like websites, apps, and artificial intelligence (AI). Inside, rows of computers called servers do most of the work by processing and storing data, and they typically use the largest share of energy (about 60% in many data centers). Other equipment, such as storage systems and networking equipment, help store information and move data between devices, using a smaller portion of energy.

To keep everything running, data centers rely on support systems. Cooling systems prevent equipment from overheating and can use a significant amount of energy—sometimes up to 30% or more. Uninterruptible power supply and backup systems ensure the data center stays online even during outages.

Source: IEA

Questions

  1. Based on the information above, which part of the data center uses the most energy? About what percentage does it use?







  1. Which system uses the second most energy, and approximately what percentage does it use?







Data Center Impacts

Data centers are critical for supporting modern technology and AI, but they require significant energy, water, and land resources. Their overall impact depends on how they are powered, where they are located, and how efficiently they are managed.

Energy Use

Data centers require large amounts of electricity to power servers and cooling systems. The amount of energy a data center uses varies greatly and depends on many factors. While many current large data centers use 100-300 megawatts (MW) of power, some of the newest AI facilities in development are being designed to use gigawatt-scale power (1,000+ MW). They could use as much electricity as a million homes, similar to a large city.

Where Does the Energy Come From?
Currently, most data centers draw power from the regional electric grid, meaning their energy comes from the mix of sources used to generate electricity in that area. Recently, some very large AI data centers in development have begun developing their own power sources, such as on-site natural gas plants, to ensure reliable electricity and avoid delays connecting to the public grid. These delays can occur because grid interconnection requires extensive studies, permitting, and infrastructure upgrades, which can take years to complete, and are increasingly slowed by large backlogs of projects waiting to connect to the grid. Some companies are also exploring nuclear energy and other dedicated power solutions for future data centers. 

Questions

  1. How is the scale of electricity use changing between current data centers and newer AI-focused data centers?







  1. How do most data centers get power today, and what new approaches are being developed?







Water Use (Cooling)

Because data centers generate large amounts of heat, they require cooling systems that often use significant amounts of water. Some systems reuse water in a closed loop, but many large facilities rely on evaporative cooling, where water is lost to the air. As a result, large data centers can consume millions of gallons of water per day, depending on their size and cooling method. 

Environmental Concerns
Water may come from municipal systems, groundwater, or nearby rivers and lakes. This demand can strain local water supplies, especially in drought-prone regions. In addition to high water use, data centers can raise environmental concerns related to water quality and local ecosystems. Many facilities use cooling towers that discharge “blowdown” water containing concentrated minerals, treatment chemicals, and sometimes trace metals, which must be carefully treated before release. While most systems are regulated and designed to minimize pollution, improper handling or accidents (such as leaks of glycol-based cooling fluids) could pose risks to local water supplies. Additionally, warm water discharge can affect aquatic ecosystems. Because of these concerns, many newer data centers are adopting water recycling, treatment, and reuse systems to reduce both water consumption and environmental impact.

Questions

  1. What is the difference between evaporative cooling and closed-loop cooling systems in how they use water?







  1. What environmental risks are associated with data center cooling systems, and how can they be reduced?







Land Use 

Modern data centers vary widely in size, ranging from small server rooms inside office buildings to massive, stand-alone campuses the size of hundreds of football fields. The largest and newest AI “hyperscale” data centers in development range from 300 to 1000+ acres, making them more similar to large industrial or energy complexes than traditional buildings. 

Where Are They Built? 
Data centers are often located near cities to reduce latency (the time it takes for data to travel), which improves responsiveness for users, or in rural areas where land is cheaper. Key factors in choosing a site include access to electricity, fiber internet, water resources, and land cost.

Questions

  1. What factors influence where data centers are built? Why are these factors important?







  1. How might building a data center that covers hundreds of acres affect land use in a community?







Air and Emissions

Data centers themselves do not directly emit carbon dioxide, but the electricity they use can result in emissions depending on how that power is generated. For example, if a large data center is powered by an on-site natural gas plant, it can contribute to both greenhouse gas emissions and local air pollution. Burning natural gas primarily releases carbon dioxide (CO2), but it also produces nitrogen oxides (NOx), which contribute to ground-level ozone and smog that can have negative impacts on air quality and human health.

In contrast, if a data center is powered by low-carbon sources such as nuclear energy, the immediate air pollution impacts on nearby communities are much lower, because these sources do not produce combustion-related air pollutants during operation. Additionally, diesel-fueled backup generators, which are commonly used for reliability, emit carbon dioxide (CO2), along with significant amounts of particulate matter (PM), often referred to as soot, and nitrogen oxides (NOx), which can contribute to local air quality concerns, especially during testing or power outages. 

Question

  1. How do on-site power systems, such as natural gas plants or diesel generators, impact local air quality?







Other Impacts: Noise and Heat

Data centers can create noise from cooling systems and backup equipment, which may affect nearby communities. They also produce large amounts of heat, though some facilities are exploring ways to reuse this waste heat.

While noise concerns have traditionally focused on cooling systems, complaints are also linked to on-site energy production, such as diesel generators and natural gas turbines. In some newer facilities, these systems operate more regularly rather than only during emergencies, making them a more persistent source of noise. 

As demand for data centers grows, many new facilities are turning to on-site power generation. This may shift their overall noise profile from primarily cooling-related sounds to something more like a small power plant, increasing the likelihood of community concerns and regulatory scrutiny. 

Question

  1. How might increased use of on-site power generation change data center noise?







Economic Impacts

When in construction, data centers can bring economic opportunities such as jobs for technicians and trade workers, and opportunities for full-time staff, along with increased local tax revenue. However, they generally create fewer long-term jobs compared to other industries. 

Question

  1. Why might data centers create fewer long-term jobs compared to other industries? 







For more information, see: