On July 10, 2024, a lightning arrestorA device that protects electrical equipment by diverting lightning strikes and electrical surges to the ground. failed on a transmission line in Northern Virginia. Within milliseconds, 60 data centers simultaneously disconnected from the grid and switched to backup generators, dropping 1,500 megawatts of load in under a second[s]. Grid operators scrambled to prevent cascading blackouts across the region. The incident, revealed months later in regulatory filings, illustrated a growing crisis: AI data center energy demand is overwhelming the infrastructure meant to deliver it.
AI Data Center Energy Demand Is Exploding
U.S. data centers consumed 183 terawatt-hoursA unit of electrical energy equal to one trillion watt-hours, used to measure large-scale electricity consumption. of electricity in 2024, more than 4% of the country’s total consumption[s]. By 2030, that figure is projected to grow by 133% to 426 terawatt-hours. A January 2026 report by Bloom Energy predicts total U.S. data center power demand will nearly double between 2025 and 2028, jumping from 80 to 150 gigawattsA unit of power equal to one billion watts, commonly used to measure the electrical capacity of large power plants or data centers.[s]. That is like adding a country with the energy needs of Spain in just three years.
The scale of individual facilities compounds the problem. Meta’s Hyperion project in Louisiana will need at least 5 gigawatts to run, three times as much electricity as the entire city of New Orleans[s]. Amazon, Microsoft, Google, and Meta collectively spent over $200 billion on capital expenditures in 2024, a 62% year-over-year increase[s]. This spending reflects a race to secure AI infrastructure before competitors, but the grid cannot keep pace.
The Grid Cannot Keep Up
Local utilities are buckling under the pressure. In Virginia, data centers consumed about 26% of the state’s total electricity supply in 2023[s]. Dominion Energy now has 40 gigawatts of contracted power for data centers as of December 2024, up 88% since July 2024[s]. An April 2025 Deloitte survey of 120 U.S. power company and data center executives found that grid stress was the leading challenge for data center infrastructure development[s].
The July 2024 Virginia incident exposed how fragile this system has become. When 60 data centers dropped off simultaneously, operators had to rapidly reduce power generation to prevent a surge that could have triggered widespread blackouts[s]. John Moura, director of reliability assessment at NERC, warned: “As these data centers get bigger and consume more energy, the grid is not designed to withstand the loss of 1,500MW data centers”[s].
Consumers Are Paying the Price
AI data center energy expansion is driving electricity costs higher for everyone. Before 2019, electricity prices had been flat at around 13 cents per kilowatt-hour for more than a decade. By the end of 2025, average U.S. electricity prices had increased to 19 cents per kilowatt-hour, about 27% higher than in 2019[s]. In Virginia, electricity prices have increased by up to 267% over the past five years[s].
Utilities requested more than $29 billion in rate increases in the first half of 2025, double the amount requested in the first half of 2024[s]. One study from Carnegie Mellon University estimates that data centers and cryptocurrency mining could lead to an 8% increase in the average U.S. electricity bill by 2030, potentially exceeding 25% in the highest-demand markets of central and northern Virginia[s].
The burden falls hardest on vulnerable communities. Total outstanding utility bill debt for U.S. households reached $25 billion in June 2025, up from about $15 billion in early 2022. Utility shut-offs skyrocketed to 3.5 million in 2024 and may have reached 4 million in 2025[s].
Why New Power Takes So Long
Companies wanting to build new data centers face extraordinary delays. There is currently a seven-year wait on some requests for connection to the grid[s]. Power constraints are extending data center construction timelines by 24 to 72 months[s]. Shortages of transformers, switchgearsElectrical equipment used to control, protect, and isolate sections of a power distribution system., and gas turbines compound the problem.
The Department of Energy reports that data center load growth has tripled over the past decade and is projected to double or triple again by 2028[s]. Deloitte estimates that by 2035, AI data center energy demand in the United States could grow more than thirtyfold, reaching 123 gigawatts, up from 4 gigawatts in 2024[s]. Building the infrastructure to meet that demand under current timelines appears nearly impossible.
What Happens Next
A November 2025 Consumer Reports survey found that 78% of Americans are somewhat or very concerned that new data centers will make their energy bills go up[s]. Some companies are exploring on-site power generation to bypass grid constraints. Others are looking at nuclear power, with plans to revive retired plants like Three Mile Island in Pennsylvania.
The paradox is stark: AI promises to optimize everything, yet its physical infrastructure is choking the grids it depends on. Until power generation and transmission can catch up with demand, AI data center energy requirements will continue to strain local utilities, raise consumer bills, and threaten grid reliability.
AI Data Center Energy: Scale of the Demand Surge
U.S. data centers consumed 183 terawatt-hoursA unit of electrical energy equal to one trillion watt-hours, used to measure large-scale electricity consumption. of electricity in 2024, representing more than 4% of total national consumption[s]. The Lawrence Berkeley National Laboratory predicts this will grow from 176 TWh in 2023 to between 325 and 580 TWh by 2028, representing 6.7% to 12% of projected national electricity consumption[s]. A January 2026 Bloom Energy report projects total U.S. data center power demand will nearly double between 2025 and 2028, from 80 to 150 gigawattsA unit of power equal to one billion watts, commonly used to measure the electrical capacity of large power plants or data centers.[s].
Individual hyperscaleA data center of extremely large scale, typically consuming 20 to 500 megawatts, operated by major tech companies for cloud computing and AI workloads. facilities are reaching unprecedented power requirements. Meta’s Hyperion campus in Louisiana will require at least 5 GW, three times the electricity consumption of New Orleans[s]. The top three hyperscalersLarge technology companies that operate cloud computing infrastructure at massive global scale, such as Meta, Google, and Amazon.’ largest U.S. data centers currently draw less than 500 MW each, but facilities under construction are targeting 2 GW capacities[s]. Amazon, Microsoft, Google, and Meta collectively spent over $200 billion on capital expenditures in 2024, a 62% year-over-year increase[s].
Grid Stress Mechanisms and Capacity MarketA forward auction in wholesale electricity markets where power producers bid to guarantee future generation availability, ensuring the grid can meet projected demand. Impacts
The concentration of AI data center energy demand in specific regions creates acute grid stress. In 2023, data centers consumed approximately 26% of Virginia’s total electricity supply, with North Dakota at 15%, Nebraska at 12%, Iowa at 11%, and Oregon at 11%[s]. Dominion Energy reported 40 GW of contracted power for data centers as of December 2024, an 88% increase since July 2024[s].
The July 2024 Northern Virginia incident revealed critical vulnerabilities. A lightning arrestorA device that protects electrical equipment by diverting lightning strikes and electrical surges to the ground. failure on a 230 kV transmission line caused voltage fluctuations that triggered 60 data centers to simultaneously switch to backup power. The 1,500 MW load drop occurred in under a second, far faster than a typical power plant failure[s]. Grid operators had to rapidly reduce generation to prevent voltage surges that could have caused cascading outages[s]. NERC launched a large loads taskforce following the incident to study the reliability implications.
Capacity market pricing reflects the supply-demand imbalance. In PJM’s electricity market spanning Illinois to North Carolina, data centers accounted for an estimated $9.3 billion price increase in the 2025-2026 capacity market auction, with residential bills expected to rise by $18 per month in western Maryland and $16 per month in Ohio as a result[s]. Clearing prices for the subsequent 2026-2027 delivery year hit a FERC-approved price cap, reflecting continued supply constraints.
Consumer Cost Transmission and Affordability Crisis
The infrastructure buildout required to serve AI data center energy loads is driving rate increases across the country. Before 2019, electricity prices had been flat at approximately 13 cents per kWh. By the end of 2025, average U.S. electricity prices reached 19 cents per kWh, a 27% increase[s]. In Virginia, prices have increased by up to 267% over five years[s]. Residential prices jumped 7.1% in 2025, more than double the inflation rate[s].
Utilities requested more than $29 billion in rate increases in the first half of 2025, double the amount requested in the same period of 2024[s]. Carnegie Mellon University researchers estimate data centers and cryptocurrency mining could increase average U.S. electricity bills by 8% by 2030, with increases exceeding 25% in central and northern Virginia[s].
The affordability crisis is concentrated among vulnerable populations. Total outstanding utility bill debt reached $25 billion in June 2025, up from approximately $15 billion in early 2022. Utility disconnections reached 3.5 million in 2024 and may have exceeded 4 million in 2025[s]. Low-income households spend up to 20% of their income on energy, compared to 3% for higher-income households.
Interconnection QueueThe utility process and waiting list for reviewing requests to connect new electrical loads to the power grid. Bottlenecks and Construction Delays
Grid connection timelines have become the primary constraint on AI data center energy expansion. There is currently a seven-year wait on some requests for grid connection[s]. Power availability constraints are extending construction timelines by 24 to 72 months[s]. The U.S. interconnection queue contains more than 2 terawatts of generation and storage capacity seeking connection, exceeding the total installed capacity of the existing power fleet.
Supply chain constraints compound the problem. Shortages of transformers, switchgearsElectrical equipment used to control, protect, and isolate sections of a power distribution system., and gas turbines are delaying projects across the country[s]. A London Economics study found the U.S. would need to purchase 90% of global semiconductor manufacturing output over five years to support all announced data center load expected online by 2030[s].
The Department of Energy reports that data center load growth has tripled over the past decade and is projected to double or triple again by 2028[s]. Deloitte estimates that by 2035, AI data center energy demand could grow more than thirtyfold, from 4 GW in 2024 to 123 GW[s]. Meeting this demand under current regulatory and infrastructure constraints appears structurally infeasible.
Emerging Responses and Structural Implications
Some operators are pursuing alternative power procurement strategies. A growing share of planned new data center capacity is being designed with behind-the-meterPower generation installed on a customer's side of the utility meter, supplying electricity without going through the main grid. on-site generation to bypass grid constraints. Companies are exploring nuclear power, with plans to revive retired plants including Three Mile Island in Pennsylvania and Duane Arnold in Iowa[s]. Several tech companies have announced purchasing agreements with nuclear power startups.
Public concern is rising. A November 2025 Consumer Reports survey of 2,146 U.S. adults found 78% are somewhat or very concerned that new data centers will increase their energy bills[s]. An April 2025 Deloitte survey of 120 power company and data center executives confirmed that grid stress was the leading challenge for infrastructure development[s].
The fundamental paradox persists: AI systems designed to optimize processes and improve efficiency are creating massive new demands on physical infrastructure that cannot scale at the same pace. Until generation, transmission, and regulatory frameworks adapt, AI data center energy requirements will continue to strain local utilities, drive consumer costs higher, and expose grid reliability vulnerabilities that the July 2024 Virginia incident previewed.
