Historical estimate192 TWh · 4.7%
- Metric
- Data-center electricity consumption and share of U.S. electricity
- Scope
- United States
- Period
- 2024
- Status
- modeled
Do data centers strain the grid and cause outages?
Large loads require careful planning. Interconnection studies, staged service, generation procurement, transmission upgrades, and flexible-load terms determine the risk of outages.
Key facts, with scope and limits.
Scenario range521–843 TWh · 9.5–15.3%
- Metric
- Data-center electricity consumption and share of U.S. electricity
- Scope
- United States; compounded low-to-high cases
- Period
- 2030
- Status
- forecast
Planning outlook13 of 23 areas
- Metric
- Bulk-power-system assessment areas with resource-adequacy challenges during the next decade
- Scope
- North America; data centers are one of several demand and supply drivers
- Period
- 2026–2035 outlook published in 2025
- Status
- forecast
Virginia connection timing12–36 months
- Metric
- Reported delay for future large-load additions when needed transmission improvements are pending
- Scope
- Dominion Energy Virginia territory
- Period
- Conditions reviewed in 2024
- Status
- observed
2030 demand spans a wide range
The chart combines LBNL’s historical estimate with compounded low-to-high scenarios. The 2030 range reflects uncertainty in workload, hardware, utilization, and deployment assumptions. Available grid supply falls outside this range.
2024 is a historical model estimate. 2030 is a scenario range, with 649 TWh as the current-trends reference. LBNL excludes cryptocurrency mining.
| Period | Low (TWh) | Historical estimate / reference case (TWh) | High (TWh) |
|---|---|---|---|
| 2024 model estimate | 192 TWh | 192 TWh | 192 TWh |
| 2030 scenarios | 521 TWh | 649 TWh | 843 TWh |
Use the national baseline to read the local record.
National baseline
LBNL's latest reference case reaches 649 TWh in 2030. The forecast does not model whether every announced load can obtain supply. Grid operators can sequence or condition connections; generation adequacy and transmission readiness still require local proof.
Local case
Building type alone does not determine reliability. For a local answer, identify the serving utility, transmission zone, capacity accreditation, upgrade schedule, and the enforceable response when supply is tight.
What the evidence supports.
A data center automatically makes local outages more likely.
Not inevitable
Utilities and grid operators study large connections and can delay service until upgrades are ready. Virginia's review found transmission planning appeared to protect reliability, even as the utility delayed future additions.
The grid has plenty of room, so no special planning is needed.
Not credible
National demand is growing quickly, and Virginia's review found generation-capacity risks remained. FERC opened proceedings across all six organized markets because the integration rules and cost allocation need work.
Ask for these local records.
Without these inputs, a project-specific verdict is incomplete. Treat missing evidence as an open question.
- 01The executed interconnection and electric-service agreements, including every phase and ramp date
- 02Transmission and distribution studies, required upgrades, contingencies, and responsible payer
- 03The regional operator's accredited-capacity outlook and the utility's generation plan
- 04Curtailment, demand-response, backup-generation, and emergency operating terms
- 05Cancellation deposits, collateral, queue milestones, and the treatment of speculative requests
Sources used on this page.
- Official reportGrade BUnited States Data Center Energy Usage Report: 2025 Update
United States; history through 2024 and scenarios through 2030
A bottom-up historical estimate and scenario forecast. The forecast does not model future grid-supply constraints or departures from announced demand. - Official reportGrade BPowering Intelligence 2026
U.S. data-center annual energy by state; 2024 low, medium, and high estimates
These modeled estimates draw on incomplete public and commercial project data rather than a census of metered load. EPRI includes cryptocurrency mining, unlike the LBNL national series. - RegulationGrade AFERC large-load integration show-cause orders
All six U.S. regional transmission organizations and independent system operators
Show-cause orders begin a federal process; they do not establish a final national interconnection rule or determine one project's reliability effect. - Official reportGrade B2025 Long-Term Reliability Assessment
23 North American bulk-power-system assessment areas
This midpoint planning outlook is assembled from regional submissions and should not be read as a prediction. Data centers are one of several demand and supply drivers. - Official reportGrade B2025 State of Reliability Overview
North American bulk-power-system performance during 2024
Covers the bulk power system, not most lower-voltage distribution interruptions experienced by customers. - AuditGrade BData Centers in Virginia
Virginia, primarily FY2021–FY2023, with selected forecasts
Virginia-specific. Several values are stakeholder estimates or model outputs, and future utility-cost scenarios are explicitly uncertain. - Official reportGrade BElectricity Rate Designs for Large Loads: 2026 Update
55 U.S. large-load tariffs available through 2026
The survey covers 55 published large-load tariffs. It cannot show that any one design will eliminate every cross-subsidy or forecast error. - RegulationGrade AData Center Initiatives factsheet
Dominion Energy Virginia qualifying large-load class and contracts
This is a prospective Virginia rate design. Its safeguards reduce risk but cannot establish the future bill effect of a particular load forecast.