US Electricity Statistics 2026
By Axis Intelligence Research
Co-author: Aidan Jad | Last updated: September 2, 2026 | License: CC BY 4.0
U.S. electricity consumption is on track for 4,267 billion kilowatthours in 2026, an increase of 73 BkWh over 2025, with data centers and manufacturing driving the growth. Residential customers pay 18.27 cents per kilowatthour this year against 9.01 cents for industrial buyers. Grid reserve margins look healthy on paper and thin under stress.
Quick Answer: What Do the 2026 U.S. Electricity Numbers Show?
The U.S. electric power system is adding supply faster than it is adding peak demand — 58.5 GW of new on-peak resources against an aggregate peak demand increase exceeding 11 GW, according to the North American Electric Reliability Corporation’s 2026 Summer Reliability Assessment. But according to Axis Intelligence Research, the median U.S. assessment area retains only 30.6% of its headline reserve margin once NERC’s extreme-demand and forced-outage adjustments are applied — and the four worst-retaining grids are the ones absorbing most of the new large-load interconnections.
Residential electricity prices are forecast to average 18.27 cents/kWh in 2026 and 18.65 cents/kWh in 2027, per the U.S. Energy Information Administration’s August 2026 Short-Term Energy Outlook. Natural gas holds a 40% generation share across 2025, 2026 and 2027 while coal slides from 17% to 15%.
Key Findings
- According to Axis Intelligence Research, New England is the only U.S. assessment area whose reserve margin turns negative under NERC’s extreme summer scenario, falling from an anticipated 14.0% to −0.9% — a Stress Reserve Retention Index (SRRI™) reading of −6.4.
- According to the U.S. Energy Information Administration’s August 2026 Short-Term Energy Outlook, U.S. electricity consumption averages 11.69 billion kWh per day in 2026 and 12.03 billion kWh per day in 2027.
- According to the North American Electric Reliability Corporation’s 2026 Summer Reliability Assessment, the bulk power system added more than 58.5 GW of on-peak resources year on year, led by 16.4 GW of on-peak solar capacity and 14.7 GW of battery storage.
- According to Lawrence Berkeley National Laboratory’s Queued Up 2026 Edition, about 8,200 projects representing 1,312 GW of generation and 749 GW of storage were actively seeking U.S. transmission interconnection at the end of 2025, with active natural gas capacity up 86% to 253 GW.
- According to Axis Intelligence Research, the U.S. residential-to-industrial electricity price ratio widens from 2.007 in 2025 to 2.061 in 2027, meaning households pay slightly more than double the industrial rate and the gap is opening, not closing.
How Much Electricity Does the United States Use in 2026?
The U.S. burns through 11.69 billion kilowatthours on an average day in 2026, rising to 12.03 billion kWh/day in 2027, according to EIA’s August 2026 Short-Term Energy Outlook. Annualized, that is roughly 4,267 BkWh in 2026 and 4,391 BkWh in 2027 — Axis Intelligence Research calculates those totals by multiplying the daily figures by 365 days, and both figures are in the downloadable dataset with that method note attached.
The more informative series is the annual increment. EIA’s chart gallery puts the year-on-year change in consumption at +99 BkWh in 2024, +85 BkWh in 2025, +73 BkWh in 2026 and +123 BkWh in 2027.
| Year | Consumption (billion kWh/day) | Annual change (BkWh) | Status | Source |
|---|---|---|---|---|
| 2024 | — | +99 | Observed | EIA STEO, August 2026 |
| 2025 | 11.49 | +85 | Observed | EIA STEO, August 2026 |
| 2026 | 11.69 | +73 | Forecast | EIA STEO, August 2026 |
| 2027 | 12.03 | +123 | Forecast | EIA STEO, August 2026 |
Aidan Jad: Read that column downward and the shape of the AI-load story changes. Growth decelerated in 2026 — 73 BkWh is the smallest increment of the four — then nearly doubles in 2027. That is not a demand curve driven by consumer behavior. That is a construction schedule. Load arrives when a substation energizes and a commissioning team finishes ride-through testing, not when a forecast says it should. EIA cut its Texas load-growth forecast for 2027 from 14% to 6% after the state announced a pause on new data center development on August 3, and one policy announcement moved eight percentage points of a state forecast. Anyone modeling this sector off a smooth compound annual growth rate is modeling the wrong thing.
Where the New Demand Comes From
EIA’s Annual Energy Outlook 2026 records U.S. electricity consumption growing at an average 2.1% per year over the last five years after roughly 15 years of flat consumption, and projects 0.9% to 1.6% annual growth through 2050.
Data center servers accounted for an estimated 7% of commercial sector electricity consumption in 2025 across all AEO2026 cases, per EIA’s May 19, 2026 analysis. By 2050, server consumption alone reaches 446 to 818 BkWh, and data center servers grow to 22%–33% of commercial building electricity use. EIA also assumes space cooling in data center floorspace is as much as 2.9 times as energy-intensive as non-data center floorspace, which is why the server number understates the facility number.
One historical marker worth keeping: commercial-sector electricity intensity exceeded its 2003 high of 14.9 kWh per square foot for the first time in 2031–2032 under AEO2026 projections. Commercial buildings in the United States spent two decades getting more efficient per square foot. Servers reversed that.
What Is the Average Price of Electricity in the US in 2026?
The U.S. residential price averages 18.27 cents/kWh in 2026, commercial 13.96 cents/kWh and industrial 9.01 cents/kWh, according to EIA’s August 2026 Short-Term Energy Outlook.
| Sector | 2025 (cents/kWh) | 2026 (cents/kWh) | 2027 (cents/kWh) | Source |
|---|---|---|---|---|
| Residential | 17.30 | 18.27 | 18.65 | EIA STEO, August 2026 |
| Commercial | 13.41 | 13.96 | 14.05 | EIA STEO, August 2026 |
| Industrial | 8.62 | 9.01 | 9.05 | EIA STEO, August 2026 |
| Residential−industrial spread (Axis calculation) | 8.68 | 9.26 | 9.60 | Axis Intelligence Research |
According to Axis Intelligence Research, the residential-to-industrial price ratio moves from 2.007 in 2025 to 2.028 in 2026 and 2.061 in 2027. The formula is deliberately trivial — residential price divided by industrial price, both taken from EIA STEO Table 2 for the same year — and that is the point. Household rates are rising 5.6% in 2026 and 2.1% in 2027; industrial rates rise 4.5% and then 0.4%. The two sectors are diverging inside a single national dataset, and the divergence is arithmetic, not interpretation.
Residential Price Growth by Year
| Year | Nominal residential price growth | Source |
|---|---|---|
| 2021 | 3.8% | EIA STEO, August 2026 |
| 2022 | 10.1% | EIA STEO, August 2026 |
| 2023 | 6.4% | EIA STEO, August 2026 |
| 2024 | 3.0% | EIA STEO, August 2026 |
| 2025 | 5.0% | EIA STEO, August 2026 |
| 2026 | 5.6% | EIA STEO, August 2026 |
| 2027 | 2.1% | EIA STEO, August 2026 |
Compounding the published 2022–2027 rates gives a cumulative nominal increase of 36.6% over six years — an Axis Intelligence Research calculation built from EIA’s rounded annual growth rates, and labeled as such in the dataset.
Aidan Jad: The 2026 number is the one that will end up in state rate-case testimony. A 5.6% residential increase in a year when the Henry Hub spot price is forecast at 3.44 dollars per million Btu, down from 3.53 the year before, is not a fuel-cost story. Delivered power generation fuel cost for gas runs 3.93 dollars per million Btu in 2026 against 2.41 for coal. Cheaper input fuel and rising retail rates in the same year means the increase is sitting somewhere other than the commodity — transmission, distribution, capacity procurement, storm hardening. The bill line that says “supply” is not where the movement is.
Which US Electricity Sources Generate the Most Power?
Natural gas supplies 40% of U.S. electricity generation in 2025, 2026 and 2027 alike, per EIA’s August 2026 Short-Term Energy Outlook. The share that moves is coal, and it moves down.
| Source | 2025 share | 2026 share | 2027 share | Source |
|---|---|---|---|---|
| Natural gas | 40% | 40% | 40% | EIA STEO, August 2026 |
| Coal | 17% | 16% | 15% | EIA STEO, August 2026 |
| Nuclear | 18% | 18% | 18% | EIA STEO, August 2026 |
| Wind | 11% | 11% | 12% | EIA STEO, August 2026 |
| Solar | 7% | 8% | 9% | EIA STEO, August 2026 |
| Conventional hydropower | 6% | 6% | 6% | EIA STEO, August 2026 |
| Other sources | 1% | 1% | 1% | EIA STEO, August 2026 |
Total electric power sector generation rose 37 BkWh, or 1.8%, in the first half of 2026 versus the first half of 2025. Within that: solar up 21%, hydropower up 9%, wind up 6%, natural gas up 2%, and coal generation down 39 BkWh or 11%. EIA forecasts gas generation rising 30 BkWh in 2026 and 44 BkWh in 2027, with coal falling another 15 BkWh in the second half of 2026 and 18 BkWh across 2027.
The hydropower figure carries a warning attached to it. EIA forecasts a 3% decrease in second-half 2026 hydro generation because of intensifying drought in the western United States, which converts a first-half tailwind into a second-half headwind inside the same calendar year.
Energy-related CO2 emissions fall 1.8% in 2026 and rise 0.5% in 2027, driven down by coal retirement and back up by gas-fired generation growth.
How Reliable Is the US Grid in 2026? The Stress Reserve Retention Index
Here is where the published numbers and the useful numbers separate.
NERC’s 2026 Summer Reliability Assessment reports that all assessment areas have adequate anticipated resources for normal summer peak conditions. It also publishes, in the same Table 3, each area’s reserve margin after adjusting for higher demand, generator outages and derates in extreme conditions. Almost nobody reads the second column against the first.
The Axis Stress Reserve Retention Index (SRRI™)
SRRI™ — Stress Reserve Retention Index — measures how much of an assessment area’s headline planning cushion survives NERC’s own extreme-conditions scenario.
SRRI = (Extreme-condition reserve margin ÷ Anticipated reserve margin) × 100
Inputs: both values come directly from Table 3 of NERC’s 2026 Summer Reliability Assessment, published May 2026. Scope: the 15 U.S. assessment areas; NERC’s Canadian and Mexican areas are excluded so the ranking reads as a U.S. reliability measure. Baseline reading: summer 2026. Sign convention: a negative SRRI means the area’s resources fall below demand in NERC’s extreme scenario. A reading of 100 would mean the extreme scenario costs the area nothing.
An outside analyst can recompute every figure below from the NERC PDF in under ten minutes. That is the intent.
| Rank | Assessment area | Anticipated reserve margin | Extreme-condition margin | Erosion (pp) | SRRI™ |
|---|---|---|---|---|---|
| 1 | NPCC-New England | 14.0% | −0.9% | 14.9 | −6.4 |
| 2 | MISO | 31.1% | 3.8% | 27.3 | 12.2 |
| 3 | MRO-SPP | 26.8% | 3.8% | 23.0 | 14.2 |
| 4 | PJM | 26.2% | 3.8% | 22.4 | 14.5 |
| 5 | NPCC-New York | 34.2% | 7.5% | 26.7 | 21.9 |
| 6 | SERC-Central | 18.9% | 5.6% | 13.3 | 29.6 |
| 7 | Texas RE-ERCOT | 67.9% | 20.4% | 47.5 | 30.0 |
| 8 | WECC-Basin | 37.6% | 11.5% | 26.1 | 30.6 |
| 9 | WECC-California | 55.6% | 17.0% | 38.6 | 30.6 |
| 10 | WECC-Northwest | 26.9% | 8.8% | 18.1 | 32.7 |
| 11 | SERC-East | 26.3% | 10.3% | 16.0 | 39.2 |
| 12 | WECC-Rocky Mountain | 51.9% | 21.5% | 30.4 | 41.4 |
| 13 | WECC-Southwest | 48.1% | 25.2% | 22.9 | 52.4 |
| 14 | SERC-Florida Peninsula | 25.8% | 13.6% | 12.2 | 52.7 |
| 15 | SERC-Southeast | 36.9% | 24.7% | 12.2 | 66.9 |
Rank 1 = least retention. Reserve margin inputs: NERC 2026 Summer Reliability Assessment, Table 3, May 2026. SRRI™ and erosion columns: Axis Intelligence Research.
According to Axis Intelligence Research, the median SRRI reading across the 15 U.S. assessment areas is 30.6 and the unweighted mean reserve margin erosion is 23.4 percentage points. In plain terms, the typical U.S. grid region loses roughly seven-tenths of its advertised planning cushion the moment NERC applies a plausible bad-day scenario.
According to Axis Intelligence Research, the four lowest-retention U.S. areas — New England, MISO, SPP and PJM — are also, with the exception of New England, the regions NERC identifies as absorbing the largest data-center-driven load growth. MISO’s peak load rose 2.5 GW year on year, driven by data center projections, and NERC expects that growth to accelerate in 2027 and beyond. PJM’s headline 26.2% anticipated reserve margin collapses to 3.8% under stress. The number utilities quote in press releases and the number that describes a July heat dome are not the same number.
Why ERCOT’s 67.9% Is Not What It Looks Like
Texas posts the highest anticipated reserve margin in North America at 67.9%, and the largest absolute erosion on the continent: 47.5 percentage points, landing at 20.4%. Its SRRI of 30.0 sits at the national median. A very large headline margin and a very large derate produce an ordinary result.
The underlying improvement is real, though. ERCOT’s probability of declaring an energy emergency alert during its highest-risk hour — hour ending 9:00 p.m. on the expected August peak day — is 0.43%, down from 3.1% in the 2025 assessment. NERC attributes the drop to 8,780 MW of battery storage added in 2025 plus 2,677 MW more by March 2026. Texas net internal demand actually fell 4.6% year on year, and total internal demand fell 2.3%, because updated forecasting now reflects how large computational loads behave at peak and how many of them can be curtailed.
Aidan Jad: That ERCOT line is the most underrated number in the assessment. Demand did not fall in Texas. The forecast fell, because the operator learned that a large fraction of its new load will get out of the way when told to. Curtailable load is a capacity resource wearing a customer’s uniform. Every region negotiating large-load interconnection agreements right now is negotiating over exactly this — whether the new 300 MW campus is firm demand or dispatchable demand — and the answer changes the reserve margin arithmetic more than another gas turbine would.
New England, the Only Negative Reading
New England’s anticipated reserve margin of 14.0% clears its 13% reference margin level by a single point, and there is a 41 MW shortfall if resource additions are excluded. The region cut projected net firm imports by 836 MW year on year, to 409 MW — against a system that typically imports around 3,000 MW at summer peak. NERC records a 2.6% year-on-year drop in anticipated resources driven almost entirely by that decline in firm transfers.
The New England Clean Energy Connect line, a ±320 kV HVdc import-only tie from Québec to Lewiston, Maine, came into service in January 2026 and can move up to 1,200 MW — but on a non-firm basis for this summer, which is why it does not appear in the firm import figure.
The Northwest’s Water Problem
WECC-Northwest draws 55% of its supply from hydropower, and Washington’s snow water equivalent stood at 52% of normal on April 1, 2026 — the date snowpack typically peaks. Net internal demand grew 4.6% while anticipated resources rose 1.5%, and expected Tier 1 resources dropped roughly 43%. The anticipated reserve margin fell from 32% to 26.9%. WECC’s probabilistic model returned 2.7 MWh of expected unserved energy and 0.1 loss-of-load hours, concentrated around 6:00 p.m. in early September.
How Much New Generating Capacity Is the US Adding?
The bulk power system added more than 58.5 GW of on-peak resources between summer 2025 and summer 2026, against a 2025 baseline of 1,115 GW of anticipated resources.
| Resource | On-peak capacity added (GW) | Source |
|---|---|---|
| Solar PV | 16.4 (from 30.5 GW nameplate) | NERC 2026 SRA |
| Battery storage | 14.7 | NERC 2026 SRA |
| Natural gas | 6.7 | NERC 2026 SRA |
| Wind | 1.6 | NERC 2026 SRA |
| Other capacity change | 19.0 | NERC 2026 SRA |
| Total | 58.5 | NERC 2026 SRA |
Against that, aggregate peak demand across all 23 assessment areas rose by more than 11 GW, exceeding the 10 GW rise that preceded summer 2025. Nineteen of 23 areas posted demand increases, the largest in the U.S. West: WECC-Southwest at +9.5%, WECC-Northwest at +4.6%, WECC-Basin at +3.8%.
According to Axis Intelligence Research, the resulting capacity absorption ratio for summer 2026 is approximately 5.3 to 1 — 58.5 GW of on-peak additions divided by the 11 GW aggregate peak demand increase. Because NERC reports demand growth as exceeding 11 GW rather than as a precise value, 5.3 is an upper bound, and it is published here as one. A ratio calculated against a floor is a ceiling, and pretending otherwise would be the kind of false precision this dataset exists to avoid.
Why Do Power Projects Take So Long to Connect to the Grid?
Because the interconnection queue is roughly twice the size of the installed fleet and most of it never gets built.
At the end of 2025, about 8,200 projects representing 1,312 GW of generation and 749 GW of storage were actively seeking transmission interconnection, per Lawrence Berkeley National Laboratory’s Queued Up 2026 Edition, compiled from all seven ISOs/RTOs and 50 non-ISO utilities covering about 98% of installed U.S. generating capacity.
| Resource in active queue | Capacity (GW) | Change during 2025 | Source |
|---|---|---|---|
| Solar | 773 | −19% | LBNL Queued Up 2026 Edition |
| Storage | 749 | −16% | LBNL Queued Up 2026 Edition |
| Natural gas | 253 | +86% | LBNL Queued Up 2026 Edition |
| Wind | 220 | −19% | LBNL Queued Up 2026 Edition |
Total active queue volume fell 10% during 2025 on high withdrawal rates and fewer new requests. Some 549 GW already holds a draft or executed interconnection agreement without reaching commercial operation, including 256 GW of solar, 161 GW of storage, 76 GW of wind and 45 GW of gas. Median duration from interconnection request to commercial operation exceeded five years for projects built in 2025. Of the capacity that submitted requests between 2000 and 2020, only 13% had reached commercial operation by the end of 2025; 75% withdrew and 10% remained active.
Aidan Jad: An 86% jump in active gas capacity in a single year, while solar, wind and storage all shrink by roughly a fifth, is the clearest signal in the entire dataset about what developers think large loads will pay for. Gas queue entries are not a forecast of gas generation — three-quarters of historical queue capacity never gets built, so the 253 GW should be read as intent, not delivery. But intent moved 86% in twelve months while everything else moved backward. That is a repricing of firmness, and it happened before a single one of those turbines poured concrete.
Which Regions Face the Highest Grid Risk This Year?
NERC assessed all areas as adequate for normal peak conditions and flagged three as elevated risk under above-normal or extreme conditions: NPCC-New England, WECC-Northwest and MRO-SaskPower. A localized constraint was also identified in Far West Texas, where transmission limits can bind during high demand combined with low wind and no solar.
Two operational events shaped NERC’s framing of large loads. The Eastern Interconnection and the ERCOT Interconnection each experienced approximately 1,500 MW of voltage-sensitive load reduction. In the Eastern event, data centers and other power electronic loads transferred to backup generation, causing frequency overshoot and high voltages. The ERCOT event involved varied loads reducing consumption during an extended low-voltage period in West Texas following a protection system misoperation. NERC issued a Level 3 essential actions alert on large loads and a reliability guideline on emerging large-load risk mitigation.
Timing matters as much as magnitude. March 2026 was the warmest March in 132 years of contiguous-U.S. recordkeeping, averaging 9.4°F above the 20th-century average — heat arriving while spring maintenance outages were still in progress. EIA forecasts 1,579 population-weighted cooling degree days in 2026 against 1,540 in 2025.
Methodology
Collection. Every figure in this article was retrieved on September 2, 2026 from a primary source document opened during production: EIA’s August 2026 Short-Term Energy Outlook (full report PDF and Tables 1 and 2), EIA’s Annual Energy Outlook 2026 landing page and its May 19, 2026 Today in Energy analysis of data center server energy use, NERC’s 2026 Summer Reliability Assessment (May 2026), Lawrence Berkeley National Laboratory’s Queued Up 2026 Edition, and EIA’s Electric Power Monthly release schedule. No figure in this article was taken from a secondary compilation or an aggregator.
Formulas.
- SRRI™ = (extreme-condition reserve margin ÷ anticipated reserve margin) × 100, using NERC 2026 SRA Table 3 values for the 15 U.S. assessment areas.
- Reserve margin erosion = anticipated reserve margin − extreme-condition reserve margin, in percentage points.
- Capacity absorption ratio = on-peak resource additions (58.5 GW) ÷ aggregate peak demand increase (11 GW).
- Residential-to-industrial price ratio = residential price ÷ industrial price, same year, EIA STEO Table 2.
- Annualized consumption = STEO daily consumption × 365.
- Cumulative residential price growth = compounded chain of EIA’s published annual growth rates for 2022 through 2027.
Scope and definitions. NERC’s anticipated reserve margin is a planning figure for normal peak conditions; its extreme-condition margin applies each area’s own stated demand, outage and derate adjustments, which differ by area — Table 3 aggregates them without harmonizing them. SRRI therefore compares each area against its own stress scenario, not against a single national scenario, and readings are informative as a ranking of retention rather than as a like-for-like severity score. EIA STEO figures for 2026 and 2027 are forecasts, marked as such throughout and in the dataset’s data_type column. Where EIA publishes rounded growth rates, chained calculations built from them inherit that rounding and are flagged in the dataset’s method_note field.
What is not published here. EIA’s Electric Power Monthly for June 2026 was released August 26, 2026, and state-level retail price data for months after June 2026 does not yet exist. Rather than project it, this dataset stops where the published record stops.
About This Dataset
us-electricity-statistics-2026.csv contains 178 rows covering U.S. electricity consumption, retail prices by sector, generation shares, reserve margins by assessment area, interconnection queue composition and the Axis SRRI™ readings. Each row carries the observation, its unit, its as-of date, the issuing organization, the source document, an absolute source URL, the retrieval date, whether the source is primary, and whether Axis Intelligence Research calculated the value. Rows flagged axis_calculated = yes carry a method note pointing to the formula disclosed above.
License: CC BY 4.0. Reuse, redistribution and adaptation are permitted with attribution.
Citation: Axis Intelligence Research, US Electricity Statistics 2026: Demand, Prices, Generation and Grid Reserve Margins, 2026, axis-intelligence.com/us-electricity-statistics/
Citation formats
APA: Axis Intelligence Research. (2026). US electricity statistics 2026: Demand, prices, generation and grid reserve margins. https://axis-intelligence.com/us-electricity-statistics/
MLA: Axis Intelligence Research. “US Electricity Statistics 2026: Demand, Prices, Generation and Grid Reserve Margins.” Axis Intelligence, 2026, axis-intelligence.com/us-electricity-statistics/.
Chicago: Axis Intelligence Research. “US Electricity Statistics 2026: Demand, Prices, Generation and Grid Reserve Margins.” Axis Intelligence, 2026. https://axis-intelligence.com/us-electricity-statistics/.
Frequently Asked Questions
Does a 26% reserve margin mean PJM can survive a July heat dome?
Not on its own. PJM’s 26.2% anticipated reserve margin describes normal peak conditions. Under NERC’s extreme scenario — higher demand plus typical and forced outages plus derates — the same system shows 3.8%, an SRRI™ reading of 14.5. PJM separately forecasts about 26% installed reserves including committed demand response against a reserve requirement margin of 18.6% needed for the one-day-in-ten-years loss-of-load criterion. Both statements are true; they answer different questions. The anticipated margin answers “is the fleet adequate,” and the extreme-condition margin answers “what happens on the worst plausible day.”
Why did EIA cut its Texas load growth forecast from 14% to 6%?
Because the Texas governor announced a pause on new data center development on August 3, 2026, to collect information on projects under review. EIA lowered its 2027 Texas electricity load growth forecast to 6% in the August 2026 Short-Term Energy Outlook, against 14% in the previous edition. The revision illustrates how much of the near-term load forecast rests on the completion timing of a countable number of individual campuses rather than on diffuse economic growth.
If 2,060 GW is sitting in the interconnection queue, why is anyone worried about capacity?
Because queue volume is not a supply pipeline. Of capacity that entered U.S. interconnection queues between 2000 and 2020, only 13% had reached commercial operation by the end of 2025 — 75% withdrew and 10% is still active, per LBNL. The more actionable figure is the 549 GW already holding a draft or executed interconnection agreement, and even late-stage projects took a median of more than five years from request to commercial operation among those built in 2025.
What is curtailable large load and why did it lower ERCOT’s peak demand forecast?
Curtailable large load is demand — usually data centers or crypto-mining facilities — that the grid operator can reduce during tight conditions under contract. ERCOT’s net internal demand for summer 2026 came in 4.6% below the prior year and total internal demand 2.3% lower, because updated modeling reflects observed large-load behavior at peak and a larger volume of load that operators can curtail to prevent emergencies. The load exists; it is simply no longer counted as firm at the peak hour.
Why are data centers a reliability concern if the grid has enough capacity?
Because of how they behave, not how much they draw. NERC recorded roughly 1,500 MW of voltage-sensitive load reduction in each of the Eastern and ERCOT Interconnections. In the Eastern event, data centers and other power electronic loads switched to backup generation, causing frequency overshoot and high voltages — a load dropping off the system stresses it much as a generator tripping does. NERC has issued a Level 3 essential actions alert on computational load modeling and a reliability guideline for emerging large loads in response.
Are electricity prices rising because of data centers?
The published data does not support a clean causal claim either way, and this dataset does not make one. What it shows: residential prices rise 5.6% in 2026 while the Henry Hub spot price falls to 3.44 dollars per million Btu from 3.53, and power-sector delivered gas costs 3.93 dollars per million Btu against coal at 2.41. Retail rates rising while input fuel prices fall points to non-fuel components — transmission, distribution, capacity and reliability spending — rather than to the commodity. Attributing that to any single load class requires utility-level rate-case data, which is not in this dataset.
Which US grid region is most exposed to drought this summer?
WECC-Northwest, where hydropower supplies 55% of the generation mix. Washington’s snow water equivalent was 52% of normal on April 1, 2026, and although total Columbia River Basin water supply was forecast near average, snowpack peaked at low elevation and melted early. Net internal demand grew 4.6% against a 1.5% increase in anticipated resources, and the anticipated reserve margin fell from 32% to 26.9%.
Is coal still meaningful in the US generation mix?
Its share falls from 17% in 2025 to 16% in 2026 and 15% in 2027, per EIA. Coal generation dropped 39 BkWh, or 11%, in the first half of 2026, with another 15 BkWh decline expected in the second half and 18 BkWh across 2027 as lower-cost gas displaces it. Six plants were held available in 2025 and 2026 under Department of Energy emergency orders issued pursuant to section 202(c) of the Federal Power Act; NERC excluded those units from anticipated resources for summer 2026.
What does the SRRI reading mean for a company siting a new facility?
It ranks how much of a region’s advertised planning cushion survives a bad day, which is a different question from whether power is available or affordable. SERC-Southeast retains 66.9% of its margin under stress and SERC-Florida Peninsula 52.7%, against 12.2% for MISO and 14.5% for PJM. A low SRRI does not predict an outage — NERC assessed every U.S. area as adequate under normal conditions — but it indicates how little slack remains before an operator reaches for emergency procedures, which is the relevant input for anyone negotiating firm versus interruptible service.