What Is Driving Data Center Electricity Costs?
Data center electricity costs refer to the price utilities and, by extension, ordinary households pay for power as massive server farms — the physical infrastructure behind cloud computing, streaming, and especially artificial intelligence — draw increasingly large amounts of electricity from the grid. Unlike a typical new factory or office building, a modern hyperscale data center can consume as much power as 800,000 homes, and because these facilities are being built at a pace utilities have never had to accommodate before, the resulting strain on grid infrastructure has become one of the most closely watched cost drivers in American electricity markets heading into 2026. Utilities across the country received requests for at least 700 gigawatts of new power connections in 2025 alone — more electricity than the entire United States consumed in all of 2023.
In the US, this surge has arrived at a moment when residential electricity prices were already climbing faster than inflation. National average residential rates rose from a stable 13 cents per kilowatt-hour for over a decade before 2019 to roughly 19 cents per kWh by the end of 2025, a 27% increase even before accounting for the newest wave of data center-driven infrastructure spending. What makes 2026 particularly significant is the emergence of hard regional evidence: some states with heavy data center concentration have seen electricity prices climb by as much as 267% over five years, while grid capacity auction prices in the country’s largest regional market surged more than 1,000% in just two years. This report breaks down exactly how much data centers are costing American ratepayers, where the impact is concentrated, and why researchers remain genuinely divided on how much of the national price increase can be attributed to data centers specifically.
Key Data Center Electricity Cost Facts in the US 2026
| Fact | Figure |
|---|---|
| National average residential electricity price (end of 2025) | 19 cents/kWh, up 27% from 2019’s 13 cents/kWh |
| Residential electricity price increase in 2025 alone | 11.5%, outpacing inflation (EIA) |
| Projected national price increase by 2030 (vs. 2025) | Up to 40% |
| Gigawatts of new power connection requests received by utilities in 2025 | 700+ GW |
| US total electricity consumption, all of 2023 (for comparison) | 477 GW |
| US data centers tracked as of 2026 | 2,000+ facilities |
| Additional data centers under construction or planned | 570+, across 40+ states |
| Total US data center electricity consumption (annual) | 176 TWh, roughly 4.4% of all US electricity |
| Projected household bill increase from data centers/crypto mining by 2030 (Carnegie Mellon) | 8% average, potentially exceeding 25% in high-demand markets |
Source: US Energy Information Administration Electric Power Monthly; Environmental and Energy Study Institute (EESI); Carnegie Mellon University research via The Network Installers; ElectricChoice.com US Data Center Power Consumption report, 2026.
Taken together, these figures describe an electricity market absorbing a genuinely unprecedented demand shock. The 700-plus gigawatts of interconnection requests utilities received in a single year exceeds the entire country’s total 2023 electricity consumption, meaning grid operators are being asked to plan for a scale of new demand that has no real precedent in the history of the modern US power grid. That scale helps explain why national residential rates, which had been essentially flat for over a decade before 2019, are now projected to climb as much as 40% higher by 2030 relative to 2025 levels.
What makes the 2026 picture more complicated than a simple “data centers are raising your bill” headline is the uneven geographic distribution of both the facilities and the price impact. With 176 TWh of annual electricity consumption concentrated in just over 2,000 tracked facilities, and Carnegie Mellon researchers projecting bill increases that could triple in high-demand markets compared to the national average, the real story of 2026 isn’t a uniform national price hike — it’s a sharply uneven one, where a relatively small number of states and utility service territories are absorbing a disproportionate share of the cost.
US Data Center Electricity Consumption Statistics 2026
| Consumption Metric | Figure |
|---|---|
| Total US data center electricity consumption (annual) | 176 TWh |
| Share of total US electricity consumption | 4.4% |
| Global data center electricity demand (2024) | 415 TWh, ~1.5% of worldwide demand |
| Global data center demand projection (2030) | 945 TWh — more than double 2024 levels |
| Cooling share of total facility energy use | 30-40% |
| Servers/IT equipment share of total facility power | 40-60% |
| Data centers by state, leading four | Texas (224), Virginia (209), California (123), Ohio (108) |
| Data centers under construction/planned nationally | 570+, across 40+ states |
| AI data centers expected to be power-supply-limited by 2027 | 40%, per industry analysts |
Source: ElectricChoice.com US Data Center Power Consumption by State, 2026; The Network Installers Data Center Energy Consumption Statistics; International Energy Agency global demand projections.
US vs Global Data Center Electricity Demand (TWh, 2024 vs 2030 Projection)
US Annual Consumption (current) ███ 176 TWh
Global Consumption (2024) █████████ 415 TWh
Global Consumption (2030 projected) ██████████████████ 945 TWh
The 4.4% share of total US electricity that data centers now consume might sound modest in isolation, but the concentration behind that figure tells a more dramatic story: Texas, Virginia, California, and Ohio alone host 664 of the country’s roughly 2,000-plus tracked facilities, meaning the actual grid impact is heavily localized rather than spread evenly across all 50 states. With 570-plus additional facilities already under construction or in active planning across more than 40 states, this concentration is likely to persist even as the industry expands geographically, since developers continue prioritizing regions with existing fiber infrastructure, favorable tax incentives, and — critically — access to cheap, reliable power.
The global trajectory adds important context for where US demand is headed: with worldwide data center electricity demand projected to more than double from 415 TWh in 2024 to 945 TWh by 2030, and roughly 40% of AI-focused data centers expected to become power-supply-limited by 2027 according to industry analysts, the fundamental constraint shaping this entire sector is shifting from land and capital availability to raw electricity access. That shift is precisely why utility interconnection queues — the waiting lists data center developers join to secure grid power — have become one of the most closely watched leading indicators for where the next wave of regional electricity cost pressure is likely to emerge.
Data Center Electricity Cost Growth and Price Increases in the US 2026
| Price Growth Metric | Figure |
|---|---|
| National residential price, pre-2019 (flat for over a decade) | ~13 cents/kWh |
| National residential price, end of 2025 | ~19 cents/kWh (27% higher than 2019) |
| Residential price increase, 2025 alone | 11.5% |
| Projected national price increase by 2030 (vs. 2025) | Up to 40% |
| Virginia electricity price increase over 5 years (high data center concentration) | Up to 267% |
| National average electricity cost increase modeled from data center/crypto growth by 2030 | 6% to 29%, varying by region |
| Regional electricity cost increase in worst-affected areas by 2030 | Up to 57% |
| CO2 emissions increase attributable to data center growth by 2030 | Up to 28% |
Source: EESI “Data Center Power Demands Are Contributing to Higher Energy Bills”; NC State News, May 2026, citing joint NC State/Carnegie Mellon/University of Pittsburgh/University of Toronto research.
National Residential Electricity Price Trajectory (Cents per kWh)
Pre-2019 (flat baseline) ████████████ ~13¢
End of 2025 █████████████████ ~19¢
2030 Projection (up to) ███████████████████████ ~26.6¢ (+40% vs 2025)
The jump from a decade-plus of flat pricing around 13 cents per kWh to nearly 19 cents by the end of 2025 marks a genuine structural break from the pricing stability American consumers had grown accustomed to since before 2019, when energy efficiency policies had actually been pushing consumption — and therefore price pressure — downward even as the broader economy grew. The 11.5% single-year increase recorded in 2025 alone outpaced general inflation, a divergence the EESI specifically flagged as contributing to a broader affordability crisis for vulnerable households already stretched thin by other rising costs.
The joint academic modeling from NC State, Carnegie Mellon, the University of Pittsburgh, and the University of Toronto, published in May 2026, offers perhaps the most rigorous regional breakdown available: a national average increase of 6% to 29% by 2030 attributable specifically to data center and cryptocurrency mining demand, with the worst-hit regions seeing costs climb as much as 57%. That research also connects the dots to environmental impact, projecting data center-driven electricity demand could push CO2 emissions up by as much as 28% relative to a counterfactual future without this growth — a reminder that the cost story and the climate story are deeply intertwined for this particular source of demand growth.
State-by-State Data Center Electricity Cost Impact 2026
| State | Data Center Count | Electricity Price Trend |
|---|---|---|
| Texas | 224 (most in the US) | Prices dropped 9.5% (2020 baseline period), ERCOT rates could rise 79% by 2027 |
| Virginia | 209 | Up 267% over 5 years in highest-impact analysis; other sources cite 15.94¢/kWh, ~10% below national average |
| California | 123 | Among highest residential rates nationally at 33.25¢/kWh (May 2026) |
| Ohio | 108 | Bills rose 12%, faster than national average, per CNBC analysis |
| Illinois | N/A (PJM territory) | Bills rose 16%, among the fastest-growing in the PJM region |
| New York | N/A | 29.93¢/kWh (May 2026), up 12.1% year-over-year |
| Hawaii | N/A | 52.00¢/kWh — highest in the nation, up 26.7% year-over-year |
| Georgia | N/A | Georgia Power proposed $15 billion generation capacity expansion to meet data center demand |
Source: Newsweek “Map Shows Electricity Costs in Every State as AI Data Centers Surge Prices,” 2026; CNBC “Electricity bills in states with the most data centers are surging,” November 2025; ElectricChoice.com state facility counts, 2026; Integrity Energy blog, April 2026.
Residential Electricity Price Growth in Select High-Data-Center States (Year-over-Year)
Illinois ████████████████ +16%
Virginia █████████████ +13%
Ohio ████████████ +12%
The state-level data reveals a genuinely contradictory picture depending on which specific metric and time frame is examined. CNBC’s November 2025 analysis found electric bills in Virginia, Illinois, and Ohio — all served primarily by the same regional grid operator, PJM Interconnection — climbed 13%, 16%, and 12% respectively, all comfortably outpacing the roughly 6% national average increase recorded over the same period. Yet other contemporaneous analyses, including one cited by RealClearEnergy, found Virginia’s average residential rate of 15.94 cents per kWh actually sits more than 10% below the national average, despite the state hosting more data centers than anywhere else in the country.
This apparent contradiction resolves once the underlying mechanism is understood: Virginia’s low average rate reflects favorable underlying generation costs and a large customer base over which to spread new infrastructure investment, while the rate of change — the pace at which prices are climbing — is what’s accelerating sharply in data-center-heavy PJM territory specifically. Georgia’s proposed $15 billion generation capacity expansion, explicitly framed by Georgia Power as a direct response to data center demand growth, illustrates how utilities are attempting to get ahead of this dynamic through massive new capital spending — spending that, in nearly every regulatory framework, eventually gets recovered at least partially through the rates charged to all customers, not just the data centers driving the need for it.
PJM Grid Capacity Auction Price Statistics 2026
| PJM Capacity Auction Metric | Figure |
|---|---|
| PJM capacity auction price, 2024/25 delivery year | $28.92/MW-day |
| PJM capacity auction price, 2026/27 delivery year | $329.17/MW-day (FERC-approved cap) |
| Percentage increase, 2024 to 2026/27 | Over 1,000% in two years |
| PJM 2024-25 auction total bill | $2.2 billion |
| PJM 2025-26 auction total bill | $14.7 billion (+500%+) |
| Data center share of 2025-26 total capacity bill | $9.3 billion (63%), per independent market monitor |
| PJM 2026/27 auction total bill (latest) | $16.1 billion (+10% from prior auction) |
| People served by the PJM grid | 65+ million, across 13 states |
| Expected customer bill impact from capacity price increase | 1.5% to 5% year-over-year, per PJM |
Source: American Action Forum, April 2026; CNBC, November 2025, citing Monitoring Analytics (PJM’s independent market monitor); PJM Interconnection public filings.
PJM Capacity Auction Total Bill Growth (Billions of Dollars)
2024/25 Delivery Year █ $2.2B
2025/26 Delivery Year ███████████████████████ $14.7B
2026/27 Delivery Year ████████████████████████████ $16.1B
The PJM Interconnection capacity auction, which secures electricity supply commitments for the largest regional grid in the country, provides perhaps the single clearest, most quantifiable data point in this entire report: prices per megawatt-day surged more than 1,000% in just two years, from $28.92 to $329.17, hitting the maximum cap federal regulators had approved. Monitoring Analytics, PJM’s own independent market monitor — not an outside advocacy group, but the entity specifically tasked with impartially assessing the market’s competitiveness — explicitly attributed 63% of the entire $14.7 billion 2025-26 capacity bill directly to data center demand, one of the most authoritative single figures available connecting data center growth to a specific, measurable cost outcome.
Because this capacity cost gets built directly into the electricity bills of the 65-plus million people served across PJM’s 13-state territory, PJM itself has projected the auction price surge will translate to a 1.5% to 5% year-over-year increase on customer bills — a relatively modest-sounding percentage that, applied against a national average bill and multiplied across tens of millions of households, represents billions of dollars in aggregate additional consumer spending flowing largely from a single grid operator’s capacity market dynamics. This kind of large-scale infrastructure cost pass-through mirrors patterns seen in other capital-intensive sectors grappling with financing pressure; our coverage of business bankruptcy statistics in the US documents a related story of how elevated costs and financing conditions have squeezed businesses across the broader economy during the same period.
Household Electricity Bill Impact Statistics 2026
| Household Bill Metric | Figure |
|---|---|
| Manassas, Virginia household bill spike (real example) | From ~$100/month to $281 in one billing cycle |
| Projected average US household bill increase by 2030 (Carnegie Mellon) | 8% |
| Projected household bill increase in high-demand markets (Carnegie Mellon) | Over 25% |
| Virginia data center share of state electricity consumption | More than 1 in 4 kWh (32 TWh of 128 TWh total, 2023) |
| Dominion Energy data center share of sales (2022) | 21% |
| New Virginia data center rate class monthly bill increase, 2026 | $11.24 (typical residential customer) |
| New Virginia data center rate class monthly bill increase, 2027 | $2.36 |
| Required contract term for new large-scale Virginia customers | 14 years, regardless of actual buildout |
Source: Consumer Reports “AI Data Centers: Big Tech’s Impact on Electric Bills, Water, and More,” March 2026; American Action Forum, April 2026; ElectricChoice.com, 2026.
Household Bill Impact Range by Market Type (Projected by 2030)
National Average Household ████████ 8%
High-Demand Market Household ████████████████████████ 25%+
The real-world case of John Steinbach, a Manassas, Virginia homeowner of nearly 40 years profiled by Consumer Reports, offers a tangible illustration of what these aggregate statistics mean at the kitchen table: his electricity bill jumped from roughly $100 to $281 in a single month, a spike he described as “so far beyond any bill that I’ve ever had.” His frustration — “they’re building them like it’s ‘Field of Dreams’… but we don’t see how that’s going to happen” — captures a sentiment increasingly common in communities near new data center developments, where local residents often feel they’re absorbing costs tied to infrastructure built primarily to serve corporate and institutional customers.
Virginia’s regulatory response, approved by the State Corporation Commission, offers one concrete policy attempt to address this exact tension: a new rate class specifically for large-scale data center customers, requiring them to sign 14-year contracts regardless of whether they ultimately build out or use the full power capacity they’ve reserved — a mechanism explicitly designed, in the SCC’s own language, to shield other ratepayers from incurring part of the rising costs. Even with that protection in place, the SCC still approved a $11.24 monthly increase for typical residential customers in 2026, illustrating that even well-intentioned cost-allocation reforms haven’t fully insulated households from the broader infrastructure buildout’s price tag. This kind of coordinated policy response to a shared infrastructure challenge echoes the type of multilateral problem-solving covered in our recap of the G20 Summit, where major economies similarly grapple with balancing rapid technological and industrial growth against the cost burden placed on ordinary citizens.
Data Center Efficiency and Power Usage Effectiveness Statistics 2026
| Efficiency Metric | Figure |
|---|---|
| Average data center PUE (Power Usage Effectiveness) | 1.56 |
| Leading hyperscale facility PUE (best-in-class) | As low as 1.09 |
| PUE definition | Total facility power ÷ IT equipment power (1.0 = perfect efficiency) |
| US data center water use for cooling (2023) | ~17 billion gallons |
| Projected water use growth by 2028 | Could double |
| Water usage effectiveness industry-wide improvement since 2021 | 39% |
| Modern closed-loop cooling water use | Under 0.2 L/kWh, vs. 20 L/kWh for traditional power generation |
| Share of total data center water footprint from the power grid (not the facility itself) | ~80%, per HyperFRAME Research |
Source: The Network Installers Data Center Energy Consumption Statistics, 2026; ElectricChoice.com water usage data; Microsoft CSR and Cloud Blog disclosures.
Data Center PUE Comparison: Industry Average vs Best-in-Class
Industry Average PUE ██████████████████ 1.56
Leading Hyperscale PUE ██████████ 1.09
Perfect Efficiency (target) █████ 1.00
The gap between the industry average PUE of 1.56 and the best-in-class figure of 1.09 achieved by leading hyperscale operators represents one of the clearest, most actionable efficiency opportunities in the entire data center cost equation. A facility operating at 1.56 PUE consumes 56% more total power than its servers alone require, largely lost to cooling and facility overhead, while a facility at 1.09 wastes only 9% — meaning the efficiency gap between an average and a leading-edge facility can represent a genuinely enormous difference in total grid demand for identical computing output.
The water consumption side of this story has followed a similarly bifurcated pattern: while total US data center water use for cooling reached roughly 17 billion gallons in 2023 and could double by 2028 on current trends, industry-wide water usage effectiveness has still improved 39% since 2021, driven by technologies like modern closed-loop cooling systems that use under 0.2 liters per kilowatt-hour, a tiny fraction of the 20 liters per kWh consumed by traditional power generation itself. This distinction matters directly for the cost conversation, since roughly 80% of a data center’s total water footprint actually originates at the power plant generating its electricity, not at the facility itself — meaning efficiency gains in electricity generation and delivery, not just in server cooling, represent an equally important lever for controlling both the environmental and cost impact of this industry’s continued growth.
The Debate: Do Data Centers Actually Raise Electricity Prices?
| Perspective | Key Finding |
|---|---|
| Mainstream/academic view (NC State, Carnegie Mellon, EESI) | Data centers driving 6-29% national increase by 2030, up to 57% regionally |
| Independent market monitor (Monitoring Analytics/PJM) | Data center demand = 63% of 2025-26 PJM capacity bill increase |
| Skeptical/industry-funded view (Institute for Energy Research) | Top 10 data center states average 14.46¢/kWh vs. 14.39¢/kWh elsewhere — statistically insignificant difference |
| Correlation coefficient cited by IER | Pearson r = -0.053 (essentially zero correlation) |
| RealClearEnergy counter-analysis | Virginia and Texas (26% of all US data centers) have below-average residential rates |
| Long-run historical context (Federal Reserve, inflation-adjusted) | Real electricity prices fell 14.6% from 1978 to 2026 despite recent increases |
| Where both sides agree | Rate of price increase (not necessarily absolute price level) is elevated in PJM-territory data center hubs |
Source: Institute for Energy Research, March 2026; RealClearEnergy, February 2026; Federal Reserve historical price data cited via RealClearEnergy.
It’s worth being direct about a genuine, unresolved debate running through this entire topic: while academic researchers, independent market monitors, and consumer advocates point to substantial evidence that data centers are driving measurable price increases — particularly the striking 63% figure from PJM’s own independent monitor — the Institute for Energy Research has published a competing analysis finding essentially zero statistical correlation between a state’s data center concentration and its electricity price level, noting that the top 10 data center states average 14.46 cents per kWh, nearly identical to the 14.39-cent average everywhere else. Their argument, echoed by RealClearEnergy’s separate analysis, is that data centers actually seek out states with already-cheap, reliable power, meaning the causality may run in the opposite direction from what headlines often suggest.
The most defensible synthesis of this genuine disagreement is that both sides are examining different things: skeptics are largely correct that the absolute level of electricity prices doesn’t correlate cleanly with data center presence, since states like Virginia and Texas retain below-average rates despite hosting more than a quarter of the nation’s facilities. But the rate of price increase — how quickly bills are climbing year over year — does appear elevated specifically within grid territories like PJM that are experiencing the fastest data center-driven demand growth, a distinction that matters enormously for how policymakers and regulators design solutions like Virginia’s new data center rate class. As this debate continues playing out in state utility commissions and academic journals throughout 2026, the underlying reality both camps agree on is that unprecedented new electricity demand is reshaping American grid economics in ways the country hasn’t had to navigate at this scale before, a challenge that intersects directly with the broader global energy transition questions explored in our coverage of the COP28 Global Climate Summit, since new fossil-fuel-backed generation capacity built to meet data center demand carries implications well beyond household bills alone.
Disclaimer: The data research report we present here is based on information found from various sources. We are not liable for any financial loss, errors, or damages of any kind that may result from the use of the information herein. We acknowledge that though we try to report accurately, we cannot verify the absolute facts of everything that has been represented.
