The demand for data centers has skyrocketed in recent years to support the exponential growth of data generation as a result of cloud computing, streaming, virtual collaboration, and especially the development of artificial intelligence (AI) platforms. These facilities are critical for innovation, economic growth, and technological competitiveness.
Data centers are also significant drivers of energy use, water consumption, and localized air quality impacts, particularly in regions experiencing rapid development. While concerns about intellectual property and creative industries are well-known, their rapid expansion also brings challenges for local communities and the environment – issues that are only now becoming clearer. Addressing these concerns will require practical solutions and cooperation across industry, government, and communities.
Strain on the Electrical Grid
Data centers require a significant amount of electricity, and new facilities and processing demands are outpacing the existing power grid capacity in the United States. To prevent outages and protect sensitive equipment, operators often rely on backup generators. In some cases, facilities operate for extended periods on on-site generation due to grid limitations, which can create additional environmental and reliability concerns.
Air Quality and Emissions Challenges
When generators become the primary power source, emissions (aka air pollution) can increase substantially compared to grid-supplied electricity, particularly when generators are used for extended periods. A recent example of this is Elon Musk’s supercomputing center, Colossus, which opened in Memphis, TN, in September 2024 and operated with 35 unpermitted temporary gas turbines. Studies of its environmental impact reached different conclusions – one university projected minimal harm based on company data, while another found satellite evidence of increased nitrogen dioxide levels nearby. This illustrates the complexity of monitoring and attributing emissions.
Similar situations have occurred elsewhere. For example, an Amazon facility in Manassas, VA, is permitted to run 97 diesel generators during outages, and Loudon County, VA – home to “Data Center Alley” – has approved permits for 176 facilities. While these centers support jobs and digital infrastructure, their cumulative emissions raise legitimate air quality concerns.
While this is an area of growing interest, estimating the impact that facilities like data centers can have on air quality is challenging given the complexity of local atmospheric conditions, and it is difficult to distinguish a single facility’s emissions from nearby sources. Pollutants can spread and change based on wind, temperature, and other weather factors, and emissions from nearby power plants, transportation, and other industries can overlap in both space and time.
Furthermore, limited site-specific monitoring and limited access to detailed emissions data makes it tough to directly link observed changes to one particular source. As a result, studies often depend on models or satellite observations, which can introduce further uncertainties and may not fully reflect short-term or localized air quality changes. These challenges often require advanced atmospheric modeling and emissions analysis to accurately assess impacts and support permitting decisions. Greater transparency and improved data collection could help communities and regulators make informed decisions.
These challenges are increasingly influencing permitting, project timelines, and community response, particularly in regions with high data center density.
Water Demand and Resource Competition
As supercomputers consume large amounts of energy, they also emit a significant amount of heat. Without effective cooling, systems can overheat, resulting in operational disruptions and data loss. Cooling supercomputers requires large volumes of water.
In 2024, a report from the U.S. Lawrence Berkeley National Laboratory estimated that data centers used around 17 billion gallons of water nationwide. In Newton County, GA, residents living near Meta’s new $750 million data center reported dry wells and construction-related dust, while the facility required nearly 500,000 gallons of water per day for cooling.
Unlike electricity, water resources are harder to expand. This creates additional challenges in drought-prone regions, where increased demand from data centers can compete directly with community needs.
Regulatory and Cost Considerations: Who Pays the Price
As data center development accelerates, states are beginning to take a more active role in evaluating resource demands and infrastructure impacts, signaling a shift toward more rigorous oversight.
Currently, data centers operate under the same regulations as other commercial or industrial facilities, even though their resource demands are far greater. In many counties, utility rates are being raised for all consumers to upgrade power grids specifically to support new data centers.
In response, several states are advancing legislation focused on transparency, cost allocation, and resource use. California, for example, has introduced and advanced legislation requiring data centers to report water use and supporting regulators’ ability to evaluate whether infrastructure costs are being subsidized by other customers. Other states, including Oregon and Minnesota, have taken steps to address how large energy users contribute to grid infrastructure costs, while states such as Iowa and others are considering requirements for water-use reporting and increased transparency.
Moving Forward with Practical Solutions
Data centers and AI are expected to continue expanding as demand increases, even as environmental and community impacts are more closely evaluated. While regulations are expected to continue developing to curb negative impacts, there are some measures that can be taken in the meantime:
- Invest in renewable energy like wind or solar power
- Replace diesel generators with natural gas generators
- Engage resource experts as early as possible in the planning phase to avoid adverse impacts
As negative impacts have outpaced regulations, cities and community members have also attempted to curb harmful impacts through litigation.
Managing the growth of data centers in a way that protects both communities and the environment will depend on clear policies, practical innovation, and open communication. As AI and cloud technologies continue to expand, cooperation between industry, regulators, and researchers will be needed to find cleaner and more efficient ways to power and cool these facilities.
Managing data center growth responsibly is a shared challenge that affects energy reliability, environmental health, and economic opportunity. By focusing on practical solutions and open communication, we can support technological progress without sacrificing community well-being.
About the Author
Dr. Alejandro Valencia joined Spheros Environmental in 2021 to support the continued growth of the company’s modeling, emissions, and transportation business areas. As a member of the Atmospheric and Emissions Modeling Group, Dr. Valencia assists with emissions and air quality modeling evaluations and environmental data analysis projects. He also conducts modeling analyses to address complex environmental problems to address clients’ research, regulatory, and litigation needs. Additionally, he builds, processes, and assesses emissions inventories to support regional-scale and near-field air quality assessments. Dr. Valencia develops applications to manipulate, visualize, analyze, and interpret observed and modeled environmental data. He is responsible for preparing technical reports and communicating results to clients and other scientific researchers. If you need technical expertise related to data centers, email Dr. Valencia at AValencia@spherosenv.com.