TL;DR
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
A November 9 MIT Technology Review report argues that electricity—not only concerns about Big Tech—is a major driver of opposition to data centers. It points to grid costs and reliability in Virginia, and to gas turbines and diesel generators planned for large off-grid projects elsewhere. The scale and local effects of proposed power systems remain a source of concern, though opposition also reflects other issues.
A November 9 report by MIT Technology Review argues that the growing backlash against data centers is driven in large part by the electricity needed to run them, rather than being only a proxy for broader unease about technology companies. The report, based on a visit to Ashburn, Virginia, says the facilities’ demands on power grids—and the pollution and infrastructure associated with supplying power—are shaping local opposition.
In Virginia’s Loudoun County, a major data center hub, many facilities connect to the existing grid. The report says an Equinix facility visited by its author receives electricity from Dominion Energy and has diesel generators for backup. It puts the average operating data center in the area at 54 megawatts of power capacity, more than the power draw of Denver International Airport. The report notes that much of Dominion’s electricity comes from natural gas.
The report describes a different approach for new projects in rural areas with little grid infrastructure: developers may build power generation onsite, often using natural-gas turbines, with diesel generators for backup. It says the smallest facilities considered attractive for AI training require about 300 megawatts, while some proposed sites could reach 6,000 megawatts. These are figures and characterizations reported by MIT Technology Review; the article does not establish that every project will reach those sizes.
In Virginia, protests have helped prompt state rules requiring data centers to cover the full cost of new transmission lines built to serve them, rather than spreading those costs among other customers, according to the report. Developers told the publication they are preparing for possible rules that could cut data centers off from grid power during peak demand. The report also cites Anthropic’s reported agreement to pay $3.4 billion for gas turbines, describing it as the largest power-generation purchase by an AI company.
Who Pays to Power Data Centers
The dispute concerns more than where buildings are placed. Large facilities can require new transmission lines or private power plants, raising questions about who pays for infrastructure, how electricity supply is affected, and what nearby communities experience. If grid upgrades are passed on broadly, households and other businesses may bear costs associated with new industrial demand; Virginia’s new cost-allocation rules address that concern by assigning new line costs to data centers, the report says.
Building generation onsite does not remove local impacts. The report notes that gas turbines and diesel generators emit air pollutants, and that private projects can bring power infrastructure to communities that may have limited influence over its location. Those effects can make data center proposals a local land-use and public-health issue as well as a debate about AI and corporate power. The report’s argument is that better community engagement or tax deals may not settle objections to the physical footprint and energy needs.
Products Worth Considering
Jackery Explorer 300 Portable Power Station with 40W Solar Panel
LG UltraGear 25G590B-B 25-Inch Full HD 1000Hz Gaming Monitor
LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz
As an affiliate, we earn on qualifying purchases.
Two Routes to Data Center Power
The report distinguishes between data centers connected to an established utility grid and newer, very large projects planned where grid capacity is scarce. Existing facilities in Loudoun County generally draw from the grid, with backup generators onsite. By contrast, some AI-focused projects elsewhere are being planned with “behind-the-meter” generation—power produced for the facility rather than supplied through the local grid.
That difference helps explain why the same public concern can take different forms. Grid-connected facilities may prompt disputes over transmission construction, electricity rates and peak demand. Onsite generation can reduce reliance on nearby grid connections, but brings fuel use, emissions and power-plant construction closer to the project. The report also describes broader sources of opposition, including fears about job losses, corporate influence and AI risks, as well as misinformation; it does not claim electricity is the only cause.
How Large Projects Will Affect Communities
The report describes planned power needs and developer expectations, but it does not establish which proposed projects will be built, how quickly they will come online, or how much electricity they will ultimately consume. The 6,000-megawatt figure is presented as the scale of an ideal facility, not as a confirmed operating data center. The article also does not quantify the effect of new demand on regional electricity prices or reliability.
It remains unclear how widely utilities will rely on fossil fuels to meet demand, whether projects will obtain enough grid access, and how often onsite generators will run. The report cites concerns about pollution and a case in Memphis in which the Environmental Protection Agency said Elon Musk violated air-quality laws while dozens of turbines powered a data center. It does not provide a full accounting of emissions across the sector. Opposition also has multiple causes, so the relative weight of electricity concerns versus other objections is not established.
Rules and Power Plans Ahead
In Virginia, data centers and utilities will have to respond to rules requiring developers to pay the full cost of new transmission lines serving their facilities. The report says developers are also preparing for possible restrictions on grid power during peak demand; whether those rules will be adopted, and how they would work, was not specified.
Elsewhere, proposed facilities will depend on decisions about grid connections, onsite generation, fuel supplies and local approvals. Communities and regulators will face questions about who bears infrastructure costs and how emissions are controlled. The report does not identify a single upcoming decision that will resolve the wider backlash. Its account suggests that power plans and local negotiations will remain central as data center projects advance.
Key Questions
What is driving opposition to data centers, according to the report?
The MIT Technology Review report argues that electricity demand and its local consequences are major drivers, including grid costs, transmission construction and pollution. It also says concerns about jobs, corporate power, AI risks and misinformation contribute.
How much power does a data center use?
The report puts the average operating data center in Loudoun County, Virginia, at 54 megawatts of power capacity. It says facilities considered attractive for AI training may require about 300 megawatts, while some ideal proposed sites could reach 6,000 megawatts. Those figures describe different facility scales, not a universal usage level.
Why would a data center build its own power supply?
Some planned facilities are in areas with little grid infrastructure, and the report says the amount of power sought may exceed what a nearby grid can provide. Developers may use onsite natural-gas turbines and backup diesel generators, though this can bring emissions and other local impacts.
Will data centers raise household electricity bills?
The report does not quantify how data center demand will affect household bills. Virginia’s rules requiring data centers to pay the full cost of new transmission lines are intended to keep those costs from being spread among other customers, but the broader effect on rates remains unclear.
Source: technologyreview.com
