Data center

History

Data centers have their roots in the huge computer rooms of the 1940s, typified by the Electronic Numerical Integrator and Computer (ENIAC), one of the earliest examples of what has become to be known as a data center. Early computer systems were complex to operate and maintain, and required a special environment in which to operate. Many cables were necessary to connect all of the components, and methods to accommodate and organize these were devised, such as standard racks to mount equipment, raised floors, and cable trays (installed overhead or under the elevated floor). A single mainframe required a great deal of power and had to be cooled to avoid overheating. Security became important— computers were expensive, and were often used for government and military purposes. As such, basic design guidelines for controlling access to the computer room were devised.

During the microcomputer industry boom of the 1980s, computers started to be deployed everywhere, in many cases with little or no care about operating requirements. However, as information technology (IT) operations started to grow in complexity, organizations grew aware of the need to control IT resources. The availability of inexpensive networking equipment, coupled with new standards for the network structured cabling, made it possible to use a hierarchical design that put servers in a specific room inside the company. The use of the term data center, as applied to specially designed computer rooms, started to gain popular recognition about this time.

During the dot-com bubble (1997–2000), there was a boom in data center construction because companies needed fast Internet connectivity and non-stop operation to establish their Internet presence. Many companies started building very large facilities, called internet data centers (IDCs), which provided enhanced capabilities, such as crossover backup: "If a Bell Atlantic line is cut, we can transfer them to ... to minimize the time of outage."

The global data center market saw steady growth in the 2010s, with a notable acceleration in the latter half of the decade. According to Gartner, worldwide data center infrastructure spending reached $200 billion in 2021, representing a 6% increase from 2020 despite economic challenges posed by the COVID-19 pandemic.

In 2024, global energy consumption from data centers was 620 TWh, with 80% of that consumption concentrated in the U.S., China, and the United Kingdom. Colocation centers were the most common category worldwide, with 4,799 colocation centers in 127 countries. In that year, the largest data center in the world was Citadel, owned by Switch, at the Tahoe Reno Industrial Center in Reno, Nevada, occupying 7.2 million square feet.

Growth of the data center industry has prompted local resistance. A 2023 survey found social mobilizations against the industry in Europe, the U.S., and South America. These movements expressed concern about the centers' energy and water use, pollution, and impact on agriculture or the landscape. A 2026 poll by Politico suggested that resistance to growth of the industry would make candidates' stance on data centers a significant issue for voters in the U.S.

In the United States

The U.S. hosted 5,381 data centers in March 2024, the highest number of any country worldwide, accounting for 40% of the global market in 2023. As of 2025, roughly one third of America's data centers were in three states: Virginia (643), Texas (395), and California (319).

The largest concentration of data centers in the world is in Northern Virginia, with over 250 facilities in the region. In 2025, Georgia was the fastest growing data center hub in the U.S., having enacted several state tax exemptions for the industry and hosting between 100 and 150 centers in the state. Local governments have responded to the growing industry in Georgia with a variety of local ordinances to guide permitting of these facilities.

In 2025, it was estimated that the U.S. GDP growth was only 0.1% without the investments in data centers for artificial intelligence.

In 2026, more than 300 jurisdictions across 44 U.S. states had enacted moratoriums on data center development, amid concerns over electricity consumption, water use, noise and local environmental impacts.

Design and architecture

Data centers house critical computing resources in a controlled environment and must generally operate with very high availability. Key design elements include providing power for the equipment, temperature and humidity control, cabling, fire safety, and security. Information security is also a concern. For this reason, a data center must provide a secure environment that minimizes the risk of a security breach.

Obsolescence and modernization

Industry research company International Data Corporation (IDC) put the average age of a data center at nine years As of August 2011. As of December 2024, the mean age of data center facilities had risen to 11.3 years, according to research analyst Jacqueline Davis of the Uptime Institute. Gartner, another research company, argued in 2010 that data centers older than seven years are obsolete. The growth in data volume (163 zettabytes by 2025) is one factor driving the need for data centers to modernize. That focus is not new: rapid obsolescence of data center equipment was a concern by at least 2007, and in May 2011, the data center research organization Uptime Institute reported that 36 percent of the large companies it surveyed expected to exhaust IT capacity within the next 18 months.

Industry standards

The Telecommunications Industry Association (TIA)'s Telecommunications Infrastructure Standard for Data Centers specifies the minimum requirements for telecommunications infrastructure of data centers and computer rooms, including single tenant enterprise data centers and multi-tenant Internet hosting data centers. The topology proposed in this document is intended to be applicable to any size data center.

Telcordia GR-3160, NEBS Requirements for Telecommunications Data Center Equipment and Spaces, provides guidelines for data center spaces within telecommunications networks, and environmental requirements for the equipment intended for installation in those spaces. These criteria were developed jointly by Telcordia and industry representatives. They may be applied to data center spaces housing data processing or IT equipment. The equipment may be used to:

• Operate and manage a carrier's telecommunication network
• Translate raw consumption data into usage metrics that integrate directly with utility billing and customer service platforms
• Continuously monitor data streams to alert utilities of irregular spikes, line leaks, or power outages.
• Provide data center-based applications directly to the carrier's customers
• Provide hosted applications for a third party to provide services to their customers
• Provide a combination of these and similar data center applications

Reliability of electrical power supply

Power supplies, either back up or continuous onsite power, consist of one or more uninterruptible power supplies, battery banks, diesel, gas turbine, and/or gas engine generating sets. Greater primary fuel energy efficiency can be achieved with the use of cogeneration technology, generating electricity, heating, and cooling onsite.

To prevent single points of failure, all elements of the electrical systems, including backup systems, are typically given redundant copies, and critical servers are connected to both the A-side and B-side power feeds. This arrangement is often made to achieve N+1 redundancy in the systems. Static transfer switches are sometimes used to ensure instantaneous switchover from one supply to the other in the event of a power failure.

Low-voltage cable routing

Options for low-voltage cable routing might include data cabling that is routed through overhead cable trays; raised floor cabling, both for security reasons and to avoid the extra cost of cooling systems over the racks; and anti-static tiles for flooring, especially in low-cost data centers.

Environmental control

Maintaining suitable temperature and humidity levels is critical to preventing equipment damage caused by overheating. Overheating can cause components (usually the silicon or copper in wires or circuits) to melt, leading to loose connections and fire hazards. Typical temperature control methods include air conditioning and indirect cooling, such as the use of outside air, indirect evaporative cooling units, and seawater cooling.

Airflow management is the practice of achieving data center cooling efficiency by preventing the recirculation of hot exhaust air and by reducing bypass airflow. Common approaches include hot-aisle/cold-aisle containment and the deployment of in-row cooling units, which position cooling directly between server racks to intercept exhaust heat before it mixes with room air.

Humidity control prevents moisture and related issues. For example, excess humidity can cause dust to adhere more readily to fan blades and heat sinks, impeding air cooling and leading to higher temperatures.

Aisle containment

Cold aisle containment is done by exposing the rear of equipment racks, while the fronts of the servers are enclosed with doors and covers. This is similar to how large-scale food companies refrigerate and store products.

Computer cabinets/server farms are often organized for containment of hot/cold aisles. Proper air duct placement prevents cold and hot air from mixing. Rows of cabinets are paired to face each other so that the cool and hot air intakes and exhausts do not mix, which would severely reduce cooling efficiency.

Alternatively, a range of underfloor panels can create efficient cold air pathways directed to the raised-floor vented tiles. Either the cold aisle or the hot aisle can be contained.

Another option is fitting cabinets with vertical exhaust duct chimneys. Hot exhaust pipes/vents/ducts can direct the air into a plenum space above a dropped ceiling and back to the cooling units or to outside vents. With this configuration, a traditional hot/cold aisle configuration is not required.

Fire protection

Data centers feature fire protection systems, including passive and active design elements, as well as implementation of fire prevention programs in operations. Smoke detectors are usually installed to provide early warning of a fire.

Although the main room usually does not allow a wet pipe sprinkler system due to the fragile nature of printed circuit boards, some systems can be used in the rest of the facility or in cold/hot aisle air circulation systems that are closed systems. These include fire sprinkler systems and misting, the latter of which uses high pressure to create extremely small water droplets and can be used in sensitive rooms because of their size.

However, there are other means to put out fires, especially in sensitive areas, usually using gaseous fire suppression, of which Halon gas was the most popular until the negative effects of producing and using it were discovered.

Security

Physical access to data centers is usually restricted. Layered security often starts with fencing, bollards, and mantraps. Video camera surveillance and permanent security guards are almost always present if the data center is large or contains sensitive information. Fingerprint recognition is becoming commonplace.

Logging access is required by some data protection regulations; some organizations tightly integrate it with access control systems. Multiple log entries can occur at the main entrance, entrances to internal rooms, and at equipment cabinets. Access control at cabinets can be integrated with intelligent power distribution units, so locks are networked through the same appliance.

Transformation

Data center transformation takes a step-by-step approach through integrated projects carried out over time. This differs from a traditional method of data center upgrades, which takes a serial and siloed approach. Typical projects within a data center transformation initiative include standardization/consolidation, virtualization, automation, and security.

Data center consolidation consists of reducing the number of data centers and avoiding server sprawl (both physical and virtual), which often includes replacing aging data center equipment. This process is aided by standardization, which ensures these systems follow a uniform set of configurations, simplifying and improving efficiency. Automating tasks such as provisioning, configuration, patching, release management, and compliance are other ways in which data centers can be upgraded. These changes are needed not just when facing a shortage of skilled IT workers. Lastly, security initiatives integrate the protection of virtual systems with existing security for physical infrastructures.

Raised floor

The first raised floor computer room was made by IBM in 1956 to allow access for wiring. During the 1970s, raised floors became more common because they allow cool air to circulate more efficiently. A raised floor standards guide (GR-2930) was developed by Telcordia Technologies, a subsidiary of Ericsson.

Lights out

The "lights-out" data center, also known as a darkened or dark data center, is a data center that has largely eliminated the need for direct access by personnel, except under extraordinary circumstances. Because staff do not need to enter the data center, it can be operated without lighting. All of the devices are accessed and managed by remote systems, with automation programs used to perform unattended operations. In addition to energy savings, reduced staffing costs, and the ability to locate the site farther from population centers, implementing a lights-out data center also reduces the threat of malicious attacks on the infrastructure.

Noise levels

Generally, local authorities prefer noise levels at data centers to be "10 dB below the existing night-time background noise level at the nearest residence."

Occupational Safety and Health Administration (OSHA) regulations require monitoring of noise levels inside data centers if noise exceeds 85 decibels. The average noise level in server areas of a data center may reach as high as 92–96 dB(A).

Residents living near data centers have described the sound as "a high-pitched whirring noise 24/7", saying "It's like being on a tarmac with an airplane engine running constantly ... Except that the airplane keeps idling and never leaves."

External sources of noise include heating, ventilation, and air conditioning equipment and energy generators.

Site selection

Location factors include proximity to power grids, telecommunications infrastructure, networking services, transportation lines, and emergency services. Other considerations should include flight paths, neighboring power drains, geological risks, and climate (associated with cooling costs).

Local political considerations, such as availability of subsidies and lack of opposition, are also important factors in locating data centers.

High availability

Various metrics exist to measure the data-availability that results from data-center availability beyond 95% uptime, with the top of the scale counting how many nines can be placed after 99%. The costs of avoiding downtime should not exceed the cost of the downtime itself.

Dynamic infrastructure

Dynamic infrastructure provides the ability to intelligently, automatically, and securely move workloads within a data center anytime, anywhere, for migrations, provisioning, to enhance performance, or build colocation centre facilities. It also facilitates performing routine maintenance on either physical or virtual systems, all while minimizing interruption. A related concept is Composable Infrastructure, which allows dynamic reconfiguration of available resources to suit needs only when needed.

Side benefits include:

• reducing cost
• facilitating business continuity and high availability
• enabling cloud and grid computing.

Network infrastructure

Communications in modern data centers are most often based on networks running the Internet protocol suite. Data centers contain a set of routers and switches that transport traffic between the servers and to the outside world, which are connected according to the data center network architecture. Redundancy of the internet connection is often provided by using two or more upstream service providers (). Some of the servers at a data center are used for running the basic internet and intranet services needed by internal users in the organization (e.g., e-mail servers, proxy servers, and DNS servers).

Network security elements are also usually deployed: firewalls, VPN gateways, intrusion detection systems, and so on. Also common are monitoring systems for the network and for some applications. Additional off-site monitoring systems are also typical in case of a communications failure inside the data center.

Energy use and environmental impact

Energy consumption and the environmental impacts it creates are a central issue for data centers. Power draw ranges from a few kilowatts (kW) for small server racks to several tens of megawatts (MW) for large facilities. Modern hyperscale data centers can exhibit power densities exceeding 100 times those of conventional office buildings, primarily due to the high concentration of servers and cooling systems required to manage continuous digital workloads. For higher power density facilities, electricity costs are a dominant operating expense and account for over 10% of the total cost of ownership (TCO) of a data center.

As of 2024, data centers in the U.S. are primarily powered by natural gas, which supplies 40% of their electricity (with renewable energy at 24%, nuclear at about 20%, and coal at about 15%). The Associated Press reported that electricity for AI data centers in the U.S. would likely come from natural gas or oil, as companies prefer using currently available power plants, which primarily use fossil fuels. Fossil energy is also often cheaper in locations where data centers are developed, and experts believe that energy demands from generative AI and data centers would be difficult to fulfill with renewable energy alone. Some companies such as Google, Amazon, and Meta have expressed interest in nuclear power for their data centers. As of 2020, according to the IEA, solar photovoltaic-generated electricity is at its lowest cost in history. Other data centers, including xAI's Colossus, OpenAI's Stargate, and Meta's Prometheus use their own off-grid natural gas plants. Electric vehicle and lithium-ion batteries have also been used for powering data centers, including for Colossus.

Power utility companies upgrade their infrastructure to handle the demands of new data centers, and the cost of these changes typically falls on residential or smaller commercial consumers. In 2025, the Mountain Valley Pipeline announced plans to expand its capacity by 25% to meet energy needs for data centers.

In December 2025, the Federal Energy Regulatory Commission (FERC) issued a unanimous order allowing data centers in the U.S. to connect directly to power plants. United States Secretary of Energy Chris Wright expressed support for un-retiring coal plants to power AI data centers. Electricity demands from AI data centers have slowed or reversed the retirement of peaking power plants in the U.S. For example, in 2025, Southern Company announced that energy use from data centers would prevent the company from retiring coal-fired power plants as it had earlier promised. In Nevada, the Desert Research Institute (DRI) calculated that 35% of the state's energy production could go to data centers by the year 2030 if all projects planned in the state as of 2026 are completed. DRI also stated, in areas where there are clusters of data centers, local consumers may end up paying the extra cost of expanding infrastructure.

Greenhouse gas emissions

In 2024, data centers were estimated to account for about 1.5% of global electricity consumption (approximately 415 TWh) and around 1% of greenhouse gas emissions according to U.S. Environmental Protection Agency (EPA). However, the rapid expansion is causing projections to rise sharply. Due to accelerated demand from AI, global data center electricity consumption is projected to more than double to around 945 TWh by 2030 in the IEA's base-case scenario, representing just under 3% of 2030 total global electricity consumption. This growing electricity demand, much of which is still generated by fossil fuels, increases the potential environmental impacts. In addition to operational energy demand, embodied life cycle emissions should be considered, which include the extraction, manufacturing, transportation, and recycling or disposal of materials used in the construction of facilities and hardware.

Global data center carbon dioxide emissions are projected to rise from an estimated 220 million tonnes in 2024 to 300–320 million tonnes by 2035. In 2024, Google and Microsoft each consumed 24 TWh – more than many countries. As a result, there is increasing industry pressure for decarbonization. Companies are pursuing direct clean energy agreements, such as Tencent who has pledged to be carbon neutral by 2030, and Microsoft's 2024 agreement to re-open the Three Mile Island nuclear power plant to provide 100% of the electric power for its AI data centers for 20 years.

Economic analysis of energy use

Energy efficiency and overhead

The most commonly used energy-efficiency metric for data centers is power usage effectiveness (PUE), calculated as the ratio of total power entering the data center to the power used by IT equipment.

PUE = ⁠Total Facility Power/IT Equipment Power⁠ = 1 + ⁠Non IT Facility Energy/IT Equipment Energy⁠

PUE measures the percentage of power used by overhead devices (cooling, lighting, etc.). The average U.S. data center has a PUE of 2.0, meaning two watts of total power (overhead + IT equipment) for every watt delivered to IT equipment. State-of-the-art data centers are estimated to have a PUE of roughly 1.2. Google publishes quarterly efficiency metrics from its data centers in operation. PUEs of as low as 1.01 have been achieved with two-phase immersion cooling.

The EPA has an Energy Star rating for standalone or large data centers. To qualify for the ecolabel, a data center must be within the top quartile in energy efficiency of all reported facilities. The Energy Efficiency Improvement Act of 2015 (U.S.) requires federal facilities—including data centers—to operate more efficiently. California's Title 24 (2014) of the California Code of Regulations mandates that every newly constructed data center have some form of airflow containment to optimize energy efficiency.

The European Union (EU) also has a similar initiative: EU Code of Conduct for Data Centres.

Efficiency improvements and renewable energy integration are helping to offset some emissions, but fossil fuels remain a major source of electricity for data center operations worldwide.

In 2011, server racks in data centers were designed for more than 25 kW, and the typical server was estimated to waste about 30% of the electricity it consumed. The energy demand for information storage systems is also rising. A high-availability data center is estimated to have a 1 MW demand and consume $20 million in electricity over its lifetime, with cooling accounting for 35% to 45% of the data center's total cost of ownership. Calculations show that in two years, the cost of powering and cooling a server could be equal to the cost of purchasing the server hardware. Research in 2018 showed that a substantial amount of energy could still be conserved by optimizing IT refresh rates and increasing server use. Research for optimizing task scheduling is also underway, with researchers looking to implement energy-efficient scheduling algorithms that could reduce energy consumption by anywhere between 6% and 44%.

The International Energy Agency has estimated that the energy efficiency of AI-related computer chips improved more than 100-fold between 2008 and 2023.

In 2011, Facebook, Rackspace, and others founded the Open Compute Project (OCP) to develop and publish open standards for greener data center computing technologies. As part of the project, Facebook published the designs of its server, which it had built for its first dedicated data center in Prineville. Making servers taller left space for more effective heat sinks and enabled the use of fans that moved more air with less energy. By not buying commercial off-the-shelf servers, energy consumption from unnecessary expansion slots on the motherboard and unneeded components, such as a graphics card, was also reduced. In 2016, Google joined the project and published the designs of its 48V DC shallow data center rack. This design had long been part of Google data centers. By eliminating the multiple transformers usually deployed in data centers, Google had achieved a 30% increase in energy efficiency. In 2017, sales for data center hardware built to OCP designs topped $1.2 billion and are expected to reach $6 billion by 2021.

Power and cooling analysis

Power is the largest recurring cost to the user of a data center. In 2008, the ASHRAE recommended a temperature range of 64.4 °F (18.0 °C) to 80.6 °F (27.0 °C), though some data centers could operate at higher temperatures. However, cooling at or below 70 °F (21 °C) wastes money and energy. Furthermore, overcooling equipment in environments with a high relative humidity can expose equipment to a high amount of moisture that facilitates the growth of salt deposits on conductive filaments in the circuitry.

A power and cooling analysis, also referred to as a thermal assessment, measures the relative temperatures in specific areas as well as the capacity of the cooling systems to handle specific ambient temperatures. A power and cooling analysis can help identify hot spots, over-cooled areas that can handle greater power use density, the breakpoint of equipment loading, the effectiveness of a raised-floor strategy, and optimal equipment positioning (such as AC units) to balance temperatures across the data center. Power cooling density is a measure of how much square footage the center can cool at maximum capacity. The cooling of data centers is the second largest power consumer after servers, with cooling taking about 7% to 30% of energy usage (depending on efficiency), compared to an average of 60% of energy used by servers.

Energy efficiency analysis

An energy efficiency analysis measures the energy use of data center IT and facilities equipment. A typical energy efficiency analysis measures factors such as a data center's Power Use Effectiveness (PUE) against industry standards, identifies mechanical and electrical sources of inefficiency, and identifies air-management metrics. However, the limitation of most current metrics and approaches is they do not include IT in the analysis. Case studies have shown that by addressing energy efficiency holistically in a data center, major efficiencies can be achieved that are not possible otherwise.

Computational fluid dynamics (CFD) analysis

Computational fluid dynamics (CFD) analysis uses sophisticated tools and techniques to understand the unique thermal conditions present in each data center—predicting the temperature, airflow, and pressure behavior of a data center to assess performance and energy consumption using numerical modeling. By predicting the effects of these environmental conditions, CFD analysis can be used to predict the impact of high-density racks mixed with low-density racks and the onward impact on cooling resources, poor infrastructure management practices, and AC failure or AC shutdown for scheduled maintenance.

Thermal zone mapping

Thermal zone mapping uses sensors and computer modeling to create a three-dimensional image of the hot and cool zones in a data center. This information can help identify optimal positioning of data center equipment. For example, critical servers might be placed in a cool zone that is serviced by redundant AC units.

Green data centers

Data centers consume large amounts of power, split between two main uses: running the equipment and cooling it. Power efficiency reduces the first category.

Cooling cost reduction through natural means includes location decisions. When the focus is on avoiding good fiber connectivity, power grid connections, and people concentrations for equipment management, a data center can be miles away from users. Mass data centers like Google or Facebook do not need to be near population centers. Arctic locations that can use outside air for cooling are becoming more popular. Countries with favorable conditions, such as Canada, Finland, Sweden, Norway, and Switzerland are trying to attract cloud computing data centers.

Singapore lifted a three-year ban on new data centers in April 2022. A major data center hub for the Asia-Pacific region, Singapore lifted its moratorium on new data center projects in 2022, granting four new projects, but rejecting more than 16 data center applications from over 20 received. Singapore's new data centers will meet very strict green technology criteria, including "Water Usage Effectiveness (WUE) of 2.0/MWh, Power Usage Effectiveness (PUE) of less than 1.3, and have a "Platinum certification under Singapore's BCA-IMDA Green Mark for New Data Centre" criteria that clearly addressed decarbonization and use of hydrogen cells or solar panels.

Energy reuse

It is very difficult to reuse the heat generated by air-cooled data centers. For this reason, data center infrastructures are more often equipped with heat pumps.

Power infrastructure

The rapid growth of artificial intelligence (AI) and high-performance computing workloads has increased attention on data center power infrastructure, particularly the relationship between electrical reliability, deployment timelines, energy costs, and decarbonization objectives. Industry publications have described phased approaches to infrastructure development that enable operators to deploy power capacity rapidly while maintaining pathways toward lower-carbon energy systems as technologies, fuel availability, grid capacity, and economics evolve.

Recent industry discussion has also emphasized "speed-to-power" as a significant consideration in data center development, reflecting the growing importance of securing electrical capacity and resilient power systems within increasingly compressed project timelines.

Morgan Stanley predicts that hi-tech firms will have invested up to $3 trillion, through 2028, building and equipping new data centers, as they compete to generate AI computing power.

An analysis by Lawrence Berkeley National Laboratory of U.S. retail electricity prices between 2019 and 2025 found that the main drivers of price increases were utility investment in grid infrastructure, natural gas prices, recovery from natural disasters and state energy policies, and did not identify data centers as a major influence on prices in most of the country. States with the largest data center demand growth generally saw price decreases over the period, although some saw increases, and it remains unclear whether the pattern will continue. A growing number of state utility regulators have adopted large-load tariffs or separate customer classes intended to assign the cost of grid upgrades required by data centers to those customers rather than to other ratepayers.

Regulation and policy responses

In the 2020s, the rapid growth of data center construction gave rise to policy debates over tax incentives, electricity grid, water resources, and climate commitments.

Tax incentives

By 2024, at least 36 U.S. states had introduced specific tax incentives to attract data center investment, most often in the form of sales tax exemptions for servers, networking equipment, and electricity. Virginia hosts the largest concentration of data centers in the U.S., and a December 2024 audit by its Joint Legislative Audit and Review Commission (JLARC) put the industry's annual contribution at roughly 74,000 jobs and $9.1 billion in gross state product. The audit reported that the state's sales and use tax exemption was claimed by about 90% of operators and that reduced state revenue by an estimated $928 million in fiscal year 2023. Thus, JLARC concluded that ending the exemption would likely shrink Virginia's share of new data center investment.

Economists have debated the effectiveness of such incentives more broadly. A review published in the Journal of Economic Perspectives (JEP) found that state and local business tax incentives in the U.S. had roughly tripled since the 1990s to about $30 billion per year, and reported no strong evidence that firm-specific incentives increase broader economic growth at the state or local level. Further, research by Timothy Bartik of the Upjohn Institute estimated that economic development incentives change firm location decisions in only 2% to 25% of cases.

In response to tax-exemption policies, several jurisdictions have begun to reevaluate incentives for data centers. In 2024, the Georgia General Assembly passed legislation to terminate the state's data center sales tax exemption. However, the bill was ultimately vetoed by Governor Brian Kemp. In another case in 2026, the Virginia Senate voted 28 to 12 to phase out the state's data center sales tax exemption as part of broader budget negotiations.

Water footprint

Most large data centers use water for evaporative cooling of their servers. A 2025 article estimated that data centers use about 0.3% of the public water supply in the United States. A separate study published in Environmental Research Letters found that about one-fifth of direct water use by U.S. data centers came from moderately or highly water-stressed watersheds, and that nearly half of the country's servers ran on electricity drawn from water-stressed regions. According to a 2024 Berkeley Lab report, data centers in the U.S. had used approximately 17 billion gallons of water. The same report expected hyperscale data centers specifically to use 16 to 33 billion gallons of water per year by 2028.

In Virginia, the December 2024 JLARC audit reported that data centers in the state consumed about 2.1 billion gallons of water in 2023, an 86% increase since 2019, and that roughly one-third of that came from reclaimed water rather than potable supplies. The audit recommended that local governments be authorized to require water-use estimates as part of the data center permitting process.

There is no federal U.S. standard requiring data center operators to disclose their water footprint, and access to facility-level data has often been limited by non-disclosure agreements (NDAs) with host municipalities. In 2022, the city of The Dalles, Oregon, settled a public records lawsuit brought by The Oregonian, after which Google ended its nationwide practice of treating site-level water use as a trade secret. In 2025, several states considered legislation to require data centers to disclose water use, including bills in Virginia, California, and New Jersey, although each was vetoed by the state's governor that year.

Electricity and grid policy

The growth of data center electricity demand in the 2020s has prompted regulatory debates over grid capacity, transmission planning, and how the costs of new infrastructure should be allocated among customers. The 2024 Lawrence Berkeley National Laboratory (LBNL) report on U.S. data center energy use, prepared at the request of the U.S. Congress, estimated that data centers accounted for about 4.4% of national electricity consumption in 2023 and projected a rise to between 6.7% and 12% by 2028, depending on growth scenarios. In 2025, the IEA projected that data centers would account for nearly half of new U.S. electricity demand growth through 2030.

The North American Electric Reliability Corporation (NERC), the federally designated electric reliability organization for the U.S., identified data centers as the single largest driver of projected demand growth in its 2024 long-term reliability assessment. Additionally, NERC documented a voltage event in the PJM Interconnection region in July 2024, in which roughly 1,500 megawatts of data center load disconnected simultaneously across about 60 facilities.

How the costs of meeting this demand are shared has also become a central policy question. The JLARC audit in 2024 estimated that, without any changes to rate design, typical residential customers of Dominion Energy Virginia could see their bills rise by $14 to $37 per month by 2040 due to data center-driven grid investment. As such, the report recommended creating a separate customer rate class for large data centers. Some researchers have urged caution in interpreting demand projections. Furthermore, a 2024 commentary in the journal Joule argued that public estimates of future data center electricity use vary widely, that utilities have historically overforecasted load growth, and called for more transparent data on actual facility-level consumption.

As data center electricity use has grown, several U.S. states have created "large-load" tariffs that require data center operators to help pay for new grid infrastructure, rather than passing those costs on to other customers. In May 2026, Oregon regulators approved a large-load tariff for Portland General Electric applying to data centers using more than 20 megawatts. The rule requires these customers to pay for the distribution upgrades needed to serve them and adds a surcharge on the largest projects, with revenue used to offset costs for lower-income customers. Not every proposal has been approved: in October 2025, the Federal Energy Regulatory Commission rejected a large-load tariff proposed by the Tri-State Generation and Transmission Association, ruling that it would have regulated retail electricity sales, which fall under state rather than federal jurisdiction. A 2026 review by researchers at Argonne National Laboratory found that U.S. data center electricity use is projected to grow between 13% and 27% annually from 2023 to 2028, far faster than overall U.S. electricity demand.

Without such tariffs, the cost of new power infrastructure can be paid by households rather than by the data centers whose demand created the need for it. The same 2026 Argonne study found that in the PJM Interconnection region, roughly four billion dollars in high-voltage transmission costs tied to data center demand were spread across all utility customers under existing planning rules. A separate 2026 review in the MIT Science Policy Review found that growing data center electricity demand can create affordability pressure for households, since infrastructure costs are often passed through in utility bills. The review noted that a single large AI-focused data center can use as much electricity as 100,000 households, with the newest facilities expected to use up to 20 times that amount. In response, several states have adopted tariffs requiring data centers to pay for the grid costs they cause, including the Oregon framework approved for Portland General Electric.

Climate policy alignment

The growth of data centers has raised questions about whether the renewable energy commitments made by their operators are consistent with broader climate goals. Many large operators report meeting electricity-related emissions targets through annual purchases of renewable energy certificates (RECs) or matching their total annual consumption with renewable generation. In 2022, a study published in Nature Climate Change analyzed 115 companies via the Science Based Targets initiative. It found that about 42% of their reported Scope 2 emissions reductions disappeared once REC-based claims were excluded. This directly suggests that some corporate climate progress reflected accounting choices rather than actual reductions in grid emissions.

In response, some operators and researchers have proposed 24/7 carbon-free energy matching. Under this approach, every hour of a facility's electricity use must be matched with carbon-free generation on the same regional grid. A 2024 Joule study modeled the idea. The authors found that annual or volumetric matching did little to cut grid emissions under current U.S. policy. Hourly 24/7 matching cut more, but it cost about $13 to $30 more per megawatt-hour.

The tension between data center expansion and corporate climate goals has also become apparent in the operators' own disclosures of information. In 2024, Google reported its greenhouse gas emissions had risen by about 48% since 2019, attributing the increase in large part to data center electricity use associated with AI workloads. Similarly, Microsoft reported in 2024 that its emissions had risen by about 29% since 2020, citing the construction and operation of new data centers. Moreover, the IEA projected in 2025 that global data center carbon dioxide emissions could peak at around 320 million tonnes by 2030 before declining slightly through 2035, with the trajectory depending heavily on the carbon intensity of the electricity supplied to new facilities.

Socioeconomic and environmental impacts

Data centers have significant impacts on local ecologies, politics, and utility infrastructure. A report from CIT and UC Riverside suggests data centers by 2030 could contribute to $20 million of healthcare costs and 1,300 premature deaths. Political and economic incentives such as tax breaks, land deals, or government subsidies drive data center growth in areas where these incentives are available. Data centers are frequently located in rural areas, where they may displace agriculture.

Growth in the data center industry has prompted several debates about environmental impacts, especially the massive energy and water demands. The supply chains, emissions, and e-waste also impose significant planetary impacts, raising questions of environmental injustice toward marginalized communities that frequently bear these burdens. Multiple studies and experts have found that data centers do not bring high-paying jobs to local communities in proportion to the tax subsidies they receive.

Water footprint

The rapid expansion of AI data centers has raised significant concerns over their water footprint, particularly in drought-prone regions. According to the IEA, a single 100-megawatt data center, using 2400 MWh of electricity per day, can use up to 2,000,000 litres (530,000 US gal) of water per day. The electricity use is equivalent to that of 60,000–120,000 households in the U.S., while water use is equivalent about 6,500 households. Water usage can be divided into three categories: on-site (direct usage from data centers), off-site (indirect usage from electricity), and supply-chain (water usage from manufacturing processes).

On-site water use refers to the direct water used by the data center for the cooling of its equipment. Water is used specifically for space humidification (adding moisture to the air), evaporative cooling systems (cooling air before entering server rooms), and cooling towers (removing heat from the facility).

Off-site water use is the indirect water usage from the electricity generated for data centers. It is estimated that 56% of U.S. data centers' electricity comes from fossil fuels in thermal power stations, which use water to generate power via steam.

Lastly, data centers consume water through the process of AI chip and server manufacturing. These chips, specifically, consume a vast amount of ultrapure water for fabrication and cooling of semiconductor plants. While this scope of water usage is not as significant as on-site and off-site water usage, it is still a contributing factor.

The water footprint of a data center depends heavily on its cooling design, local climate and electricity supply. Evaporative cooling uses less electricity but consumes water on site, whereas air-cooled and closed-loop systems consume little water directly but require more power, which increases the water consumed at the thermoelectric power plants supplying the grid. The International Energy Agency has estimated that a 100 MW data center in the United States consumes roughly 2 million litres of water per day when averaged across cooling types, of which less than 40% is consumed on site and the remainder in electricity generation. A 2026 modelling study by researchers at the University of Michigan found that switching a hypothetical facility from water-cooled to air-cooled chillers reduced total water consumption by 44%, but shifted most of the remaining footprint to power plants elsewhere on the grid. A 2026 review in Next Sustainability concluded that direct liquid cooling offers the most favourable published combination of energy and water performance, while noting that most supporting figures come from simulations and vendor data rather than independent field measurements.

Globally, since 2022, more than two-thirds of new data centers have been built in water-stressed areas, including Texas, Arizona, Saudi Arabia, and India, where freshwater scarcity is already a critical issue. The global water footprint of data centers is estimated at 560 billion litres (150×10⁹ US gal) annually, a figure projected to double by 2030 due to increasing AI demand.

In regions like Aragon, Spain, Amazon's planned data centers are licensed to withdraw 755,720 cubic metres (612.67 acre⋅ft) of water per year, sparking conflicts with farmers who rely on the same dwindling supplies. Similar tensions have arisen in Chile, the Netherlands, and Uruguay, where communities protest the diversion of water for tech infrastructure.

Tech companies, including Microsoft, Google, and Amazon, have pledged to become "water positive" by 2030, aiming to replenish more water than they consume. However, critics argue that such commitments often rely on water offsetting, which does not address acute local shortages. In 2026, Schneider Electric published modelling indicating that raising the supply temperature of liquid-cooling loops from 32°C to 45°C could reduce on-site water use by about 66% in Dallas and 82% in Paris, at the cost of additional dry-cooler capacity and capital expenditure, and only where server hardware is rated for the higher temperatures.

Trade association Data Center Coalition has refuted this image of data centers having a large water use footprint by claiming that data centers are efficient water users, often using air-based cooling 90% of the year and then switching to water-based cooling during the hottest days.

At least 59 additional data centers are planned for water-stressed U.S. regions by 2028. Kathryn Sorensen questioned the data center build out based on the water footprint.

Electronic waste

Data centers generate significant electronic waste (e-waste) due to the frequent replacement of hardware such as servers, GPUs, CPUs, memory, and storage devices, often every 2–5 years to meet demands for digital transformation and AI. Globally, e-waste totaled 62 million metric tons in 2022, with generative AI projected to contribute 1.2 to 5 million metric tons annually by 2030, including valuable metals like copper and gold, and hazardous substances such as lead and mercury. In the U.S., data centers contribute to an annual loss of $10 billion in discarded e-waste value, including $4 billion in precious metals. Only 22% of global e-waste is formally collected and recycled, exacerbating environmental pollution and health risks in informal processing sites, often in developing countries where exported waste is handled. From a political ecology perspective, data center e-waste highlights power imbalances in resource management and environmental justice, as tech corporations benefit from tax incentives while externalizing costs to marginalized communities through toxic exports and landfill burdens. This ties into debates on the commons, where shared resources like metals are depleted without equitable governance. Mitigation strategies include modular hardware designs, secure data erasure for reuse, and extended producer responsibility policies to reduce waste by up to 86% in optimized scenarios.

Impact on electricity prices

According to one 2025 energy model, the U.S. could see energy prices increase 8% nationally by 2030. Higher electricity prices are already occurring in some regions, particularly those with lots of data centers like Santa Clara, California, and parts of upstate New York. Data centers have also generated concerns in Northern Virginia about whether residents will have to foot the bill for future power lines. It has also made it harder to develop housing in London. A Bank of America Institute report in July 2024 found that increased demand for electricity due in part to AI has been pushing electricity prices higher and is a significant contributor to electricity inflation. A Harvard Law School report in March 2025 found that because utilities are increasingly in competition to attract data center contracts from big tech companies, they are likely hiding subsidies to those trillion-dollar companies in power prices by raising costs for American consumers.

Per the trade association, Data Center Coalition, not all utilities are increasing prices; some of them, such as in Indiana and Georgia, may be lowering or freezing their power rates because of data center revenue.

Research by the Electric Power Research Institute (EPRI) has examined the conditions under which new large-load demand can lower rather than raise average electricity prices. According to EPRI's analysis, new demand lowers average prices when incremental serving costs fall below the system's existing average cost, achieved either by spreading fixed infrastructure costs across a greater volume of electricity sales, or by adding generation capacity at a lower cost than the incumbent fleet. A subsequent causal study using instrumental variable estimation found that, on average, a 10 percent increase in data center capacity was associated with a 0.4 percent decline in average residential electricity rates in the United States between 2015 and 2024. Research from MIT's Center for Energy and Environmental Policy Research found that workload flexibility reduces operating costs for data center operators by between 2 and 5 percent depending on the regional grid. Beyond rate effects, data centers can provide ancillary benefits to local energy systems through on-site power generation and the repurposing of waste heat for district heating or industrial processes. In the European Union, the European Commission has projected data center energy consumption could reach 115 TWh annually by 2030 and has introduced mandatory reporting requirements and an energy efficiency labeling framework for large facilities.

In political ecology

Data centers have increasingly been analyzed through the lens of political ecology, which explores the intersections of power, politics, and environmental change in technological infrastructure. Scholars argue that these facilities reshape urban and rural landscapes by existing in marginalized or post-industrial areas, often repurposing abandoned sites while promising economic revitalization through job creation and tax revenues. For instance, hyperscale data centers in rural towns like Prineville, Oregon, or Luleå, Sweden, challenge traditional core-periphery dynamics but can exacerbate global inequalities in digital access and resource distribution. In the context of digital transformation, data centers enable the expansion of cloud computing, AI, and machine learning, but at significant ecological and social costs. Critics from science and technology studies (STS) highlight how corporate sustainability pledges, such as carbon neutrality targets by companies like Microsoft and Google, often rely on offsets and renewable subsidies that privatize benefits while causing social and environmental harms. This ties into debates on the commons, where data centers' intensive use of public resources (like electricity grids and water supplies) represents an enclosure, with tech firms receiving tax breaks and priority access at the expense of local communities.

Economic

Jobs created by data centers vary significantly by project type, phase and region.. A study found the productivity increases enabled by data centers are multiple times larger than gross external damages. Construction phases generate larger but temporary workforces, while operational roles tend to be fewer but more specialized. JLARC audit of Virginia's data center industry reported approximately 74,000 direct and indirect jobs and $9.1 billion in gross state product attributable to the sector Critics nonetheless argue that permanent employment figures are low relative to the scale of tax subsidies received, a tension that has driven several states to reassess their incentive structures.

In 2026, it was reported that data centers were becoming a target of organized crime, in which thieves were stealing copper and data center equipment from construction sites and selling them on the black market. In July 2026, authorities in Illinois recovered two trailers carrying $1.3 million worth of stolen equipment that were linked to theft cases in Alabama, Florida, and Wisconsin.

A 2026 systematic review concluded that data centers are a capital-intensive establishment with little labor demand that yield limited economic benefits to local communities while simultaneously increasing pressures on resources and infrastructure.

Opposition

Social mobilizations against data centers help shape the discourse around growth of the industry, raise public awareness of issues data centers present, and may impact public policy. Groups opposed to new data center projects have expressed concern that digital infrastructure exploits local resources 'to the benefit of actors and institutions residing elsewhere'. This may occur when data being processed in a facility belongs to actors in other countries, but local people bear the burdens of water and energy use.

In the U.S., bipartisan community resistance to data center development has grown, with residents voicing concerns about water scarcity, rising utility bills, noise pollution, and urban sprawl. Facilities in rural areas of the U.S. have been focal points for local resistance. Data Center Watch reported that multiple projects collectively worth $64 billion had been stopped or delayed between May 2024 and March 2025. An additional $98 billion worth of projects were blocked or delayed between March and June 2025. A Gallup poll taken in March 2026 reported seven out of ten Americans oppose data center projects in their area, and the most common concerns were impact on resources, electricity costs and quality of life. In August 2026, the National Republican Senatorial Committee issued a memo to AI companies to do whatever it takes to quell the public backlash against data centers, or it could "become a sleeper issue for the entire election cycle."

Outside of the U.S., protests in the Netherlands led to a temporary national ban on new mega-centers in 2022. In Chile, protests by activists to protect water reserves led to Google withdrawing plans to build a data center. In the United Kingdom, "March Against the Machines" protests were held.

Residents have been concerned about air, water, and noise pollution, as well as property devaluation, traffic, and the risk of fires. Other environmental concerns involving AI data centers include e-waste and the embodied energy of construction materials, such as concrete and cement.

Foreign propaganda

China, Russia, and Iran have spread propaganda against Western data centers in videos, comics, and news articles.

Future

Housing data centers in space has been proposed for low Earth orbit (LEO). The theoretical advantages are that of space-based solar power, in addition to aiding in weather forecasting and weather prediction computation from weather satellites, and the ability to freely scale up. However, it is difficult to design cooling systems in space, as convection is not available and sunlight causes large temperature fluctuations. Other challenges include damage from cosmic rays and micrometeorites.

Beyond questions of technical feasibility, expanding data centers presence in space has been described as the application of colonial ideology, which considers "foreign land" as readily available for resources extraction.

The design and shape of data centers could change rapidly in the not-too-distant future with the advent of new technologies, such as quantum computing, biocomputation, and synthetic DNA data storage, as examples.