U.S. Data Centre Power Costs, Grid Availability and Interconnection Pressure by State

Executive Summary: Cheap Power Is Not the Same as Available Power

The U.S. data-centre market is entering a period in which electricity cost alone is becoming an insufficient measure of site attractiveness. A state may offer relatively low industrial electricity prices, substantial generation resources or strong renewable potential, yet still present challenges for a hyperscale data-centre developer if the required load cannot be connected and served within the project's development timeline. Conversely, a higher-cost state can remain competitive when it offers credible access to transmission and distribution infrastructure, established large-load procedures, and greater visibility on connection requirements and timing. This creates the central NMSC framework: data-centre site economics = electricity cost + power availability + connection speed + grid-development requirements.

The demand-side pressure is already visible in national electricity data. EIA expects U.S. electricity sales to reach approximately 4,135 billion kWh in 2026, nearly 2% above 2025, and 4,211 billion kWh in 2027, identifying data-centre development and increased manufacturing activity as important drivers of growth. Virginia provides a particularly clear state-level example: EIA reports the state's commercial electricity sales increased by nearly 30 million MWh between 2019 and 2025, with summer peak load in the Dominion Virginia Power zone approximately 23% above 2019 levels.

At the same time, the U.S. power system has a substantial pipeline of potential new generation and storage. LBNL's 2026 Queued Up analysis identifies approximately 8,200 active projects seeking interconnection, representing 1,312 GW of generation and 749 GW of storage — but projects reaching commercial operation in 2025 had a median interconnection-to-operation duration of more than five years. Potential electricity supply is therefore not the same as immediately deliverable power for a new data-centre load.

Why This Question Is Live in 2026

The U.S. data-centre power market is moving from a relatively simple question of how much electricity is available to a more complex question of where, when and under what conditions additional electricity can actually be delivered to large new loads. Developments during 2026 highlight this shift.

The U.S. Data Centre Power-Cost Spectrum

Electricity cost provides the first layer of NMSC's assessment of U.S. data-centre location attractiveness. Because data centres are large and continuously operating electricity consumers, NMSC uses industrial electricity prices as the primary cost-ranking benchmark, with commercial prices as a complementary benchmark. The latest EIA monthly observation used in this analysis is June 2026; these figures are state-level electricity-cost benchmarks, not data-centre-specific tariffs.

The data show substantial dispersion in electricity costs across states — from 6.8¢/kWh in Texas to 20.7¢/kWh in California. Lower industrial electricity prices can improve the operating-cost case for power-intensive data centres, while higher-cost markets may carry a greater recurring electricity-cost burden. However, low electricity prices alone do not establish data-centre power attractiveness — the availability of additional generation, the ability to interconnect large new loads, transmission and distribution requirements, and the time required to deliver power are equally important considerations. The complete NMSC dataset extends this comparison across all 50 states and the District of Columbia — see the State Dataset Explorer below.

The 4 Pillars of Data Centre Power Availability

NMSC's assessment moves beyond electricity price to examine whether power can actually be secured, connected and delivered at the scale and speed required by data-centre developers. Together, these four pillars distinguish between a state that has cheap electricity and a state that has power that is economically attractive, sufficiently available and realistically deliverable to new data-centre projects.

Power-Cost and Grid-Pressure Matrix

Electricity cost and power availability represent two different considerations in data-centre site selection. A state may offer relatively low industrial electricity prices while facing increasing pressure from data-centre load growth, interconnection requirements or grid infrastructure constraints; conversely, a higher-cost market may provide a more workable pathway for securing incremental power. The matrix below plots the eight case-study states on industrial electricity cost (x-axis) against active data-centre pipeline count as a grid/large-load pressure proxy (y-axis).

This is intended as an analytical framework rather than a formal state ranking. NMSC does not assign an arbitrary 0–100 power-availability score, because the underlying indicators are derived from different sources and do not provide a standardized national measure of spare grid capacity. The framework instead highlights the central distinction in the U.S. data-centre power market: low electricity cost does not necessarily mean that new power can be obtained quickly or economically.

This produces four distinct but connected market conditions: low-cost power markets (competitive prices supporting favorable operating economics), high-growth power markets (strong data-centre development increasing electricity demand), grid-pressure markets (rapidly increasing demand requiring additional generation, transmission or distribution investment), and power-ready markets (competitive economics supported by a stronger pathway for obtaining and delivering power). Texas and Virginia sit in the high-pressure zone despite very different cost positions; Louisiana and Washington combine low cost with comparatively low pipeline pressure; California combines high cost with moderate pressure.

State-Level Case Studies

Rather than applying the full framework to all 50 states in narrative form, NMSC selects eight contrasting markets to illustrate how electricity cost, data-centre development, grid conditions and power-delivery considerations interact at the state level. The full 51-state/DC dataset is available in the State Dataset Explorer below.

These case studies demonstrate that no single factor determines data-centre power attractiveness. Virginia and Texas illustrate how strong data-centre demand can create increasing power-delivery challenges even where electricity-cost or market fundamentals are attractive. California demonstrates the impact of higher electricity costs, while Oregon and Washington illustrate markets where comparatively competitive electricity economics coexist with established data-centre development. Arizona, Louisiana and Ohio provide examples of markets where continued expansion could increase the importance of power availability, grid capacity and infrastructure requirements. The case studies are comparative market examples, not definitive statewide power-availability rankings.

Methodology and Data Limitations

The NMSC analysis evaluates U.S. data-centre power attractiveness through a combination of electricity-cost indicators, data-centre development activity, estimated power requirements and broader power-system considerations. The objective is to assess the factors that influence the ability of developers and operators to secure economically viable and timely power, rather than to produce a definitive ranking of available grid capacity by state.

Methodology

• Electricity-cost comparisons are based primarily on EIA electricity-price data; industrial prices are the primary cost-comparison benchmark, with commercial prices as a complementary perspective.
• The broader NMSC dataset combines electricity-price indicators with data-centre development indicators — operational facilities, facilities under construction, announced projects, active pipeline activity, estimated power capacity and total data-centre space — plus grid/market-region context.
• The analysis considers evidence of large-load growth, interconnection requirements, generation development and transmission/distribution infrastructure as indicators of potential power-delivery pressure, interpreted together rather than as independent measures of available electricity.
• State-level case studies illustrate different combinations of electricity economics, data-centre growth and power-system considerations, and are contrasting examples rather than a complete ranking of U.S. states.

Data Limitations

• EIA electricity prices represent average prices paid by ultimate customers, not data-centre-specific utility tariffs; actual project costs can differ substantially by utility territory, customer classification, tariff structure, demand charges, voltage level, contract terms and power-procurement arrangements.
• Data-centre facility counts, pipeline activity and estimated power-capacity figures are derived from market and industry datasets and may include projects at different development stages; announced or planned capacity should not be read as committed or secured capacity.
• Estimated data-centre power capacity represents the potential requirement associated with identified facilities — it is not spare grid capacity, immediately deliverable power, or a confirmed utility commitment.
• Interconnection queues identify projects seeking or progressing through connection processes; queued generation capacity is not equivalent to immediately available power for a new data centre.
• Power availability is highly location-specific — state-level indicators can mask significant differences between utility territories, substations, transmission zones and individual sites. This framework is a market-level screening and comparative tool, not a substitute for site-specific power-availability studies, utility discussions or interconnection assessments.
• Data sources differ in reporting periods, definitions and update schedules; NMSC identifies the relevant source and period for each figure rather than treating indicators from different datasets as directly equivalent.

Key Findings

State Dataset Explorer

Filter the full 51-state/DC dataset compiled for this report, combining electricity-cost benchmarks with data-centre facility and capacity indicators.

0 records shown

Sources: EIA Electric Power Monthly (June 2026 prices); NMSC-compiled data-centre facility and capacity indicators, Aterio data-centre development dataset (early September 2026 snapshot) — see Methodology for definitions and limitations.

Conclusion

Electricity cost remains an important input to data-centre site selection, but it is no longer sufficient on its own to identify the most attractive U.S. markets. As data-centre demand continues to grow, the ability to secure, interconnect and deliver power within a workable timeline is becoming an equally decisive factor. States such as Virginia and Texas illustrate how rapid data-centre expansion can place real pressure on generation, transmission and distribution systems, even where underlying electricity economics are attractive. States such as Oregon, Washington and Louisiana show that comparatively competitive electricity costs can coexist with more moderate development pressure — for now.

The direction of travel is clear: national demand is rising, interconnection queues remain slow and heavily subscribed, and federal and state regulators are actively rewriting the rules for how large loads connect to the grid. Developers, utilities and policymakers who treat electricity price as the only site-selection variable risk underestimating the time, infrastructure investment and regulatory engagement required to actually deliver power to a new data centre. The NMSC framework — cost, load growth, interconnection and speed to power — is intended as a starting point for that more complete assessment, not a final verdict on any single state.