USA · snapshot 2026-10-19 00:22 UTC · data to 2026-09-01
US electricity demand and grid capacity
Is US electricity demand outrunning the capacity being built to serve it?
US electricity net generation was 399,671.29 million kilowatthours in 2026-06-01, up 1.6 percent over one year, a move at the 43rd percentile of its one-year changes.
Section 1 of 9
What changed?
AI narrative Modeled
US electricity net generation reached 399671.285 million kilowatthours on 2026-06-01, a 1.6 percent rise from 393284.833 million kilowatthours one year earlier at the 43rd percentile of one-year changes [obs:410783] [obs:245385]. Net summer capacity stood at 1310.444 million kilowatts, up 4.7 percent from 1251.127 million kilowatts [obs:410776] [obs:241156]. Capacity additions were 2849.2 MW in 2026-08-01, down 39.9 percent from the prior year [obs:410638] [obs:396957]. Industrial production index rose 2.2 percent to 103.9506 [obs:415546] [obs:171164].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
US electricity net generation
Net generation is the closest stored proxy for delivered demand.
399,671.29 million kilowatthours
2026-06-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 354,690.53 (2026-05-01) | 44,980.76 (+12.7%) | 91st |
| 1 year | 393,284.83 (2025-06-01) | 6,386.45 (+1.6%) | 43rd of 629 |
| 5 years | 373,856.48 (2021-06-01) | 25,814.81 (+6.9%) | 62nd of 581 |
| 10 years | 367,903.54 (2016-06-01) | 31,767.74 (+8.6%) | 37th of 521 |
US electric net summer capacity
Installed capacity available at summer peak.
1,310.44 million kilowatts
2026-06-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 1,304.57 (2026-05-01) | 5.88 (+0.5%) | 85th |
| 1 year | 1,251.13 (2025-06-01) | 59.32 (+4.7%) | 100th of 255 |
| 5 years | 1,125.22 (2021-06-01) | 185.23 (+16.5%) | 100th of 251 |
| 10 years | 1,064.4 (2016-06-01) | 246.04 (+23.1%) | 96th of 246 |
US industrial production index
Industrial output is the largest swing load.
103.95 index 2017=100
2026-09-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 103.27 (2026-08-01) | 0.6842 (+0.7%) | 91st |
| 1 year | 101.67 (2025-09-01) | 2.28 (+2.2%) | 64th of 1280 |
| 5 years | 98.86 (2021-09-01) | 5.09 (+5.2%) | 57th of 1232 |
| 10 years | 98.75 (2016-09-01) | 5.2 (+5.3%) | 30th of 1172 |
US generation capacity additions
Capacity that actually came online, from EIA-860M.
2,849.2 MW
2026-08-01
| Span | From | Change | Percentile |
|---|---|---|---|
| 1 period | 2,984.6 (2026-07-01) | -135.4 (-4.5%) | 37th |
| 1 year | 4,738.6 (2025-08-01) | -1,889.4 (-39.9%) | 90th of 1367 |
| 5 years | 2,641.2 (2021-08-01) | 208 (+7.9%) | 43rd of 1366 |
| 10 years | 989 (2016-08-01) | 1,860.2 (+188.1%) | 84th of 1359 |
Related insights: US industrial production index (United States) rose 0.66% in 2026-09-01
Section 2 of 9
How unusual is it?
AI narrative Modeled
The latest net generation value sits at the 95th percentile of the series history, with the period change at the 91st percentile of past changes and a robust z-score of 1.6157 for the latest move; four periods have passed since a larger move [obs:245386] [obs:244759].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
- Latest value percentile
- 95th of 641 values
- Latest change percentile
- 91st of 640 changes
- Robust z of latest change
- 1.62
- Largest move
- 4 periods since a larger move
- Record level
- no
- Previous comparable move
- 2026-02-01: -56,728.57 million kilowatthours (obs 245393, 245392)
- Prior maximum
- 446,319.53 (2025-07-01, obs 245386)
- Prior minimum
- 139,589.44 (1973-04-01, obs 244759)
Section 3 of 9
Why might it be happening?
AI narrative Modeled
Documented mechanisms include the engineering relation that more delivered electricity requires more transformers needing copper, aluminum and electrical steel [mechanism:7b69c9e2-c9b8-4af3-b743-9e249d7353d5]. Statistical model drivers are not separately labeled. Hypotheses state that electricity access enables economic output and human health improvements [rel:e401fb30-2d25-4455-adf6-5b1a95069d34] [rel:f913d900-026e-4232-b9aa-a617096de9f8]. Unknowns are that data-center load growth is announced not metered and interconnection queue withdrawal rates are not ingested.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
Documented mechanisms
- More delivered electricity needs more transformers, which need copper, aluminum and electrical steelEstablished relationelectricity-system → power-grid → power-transformer → copper
Model-supported drivers
No active forecast for this topic records driver contributions (the champion may be a baseline with no drivers).
Hypotheses
- Electricity system enables Economic output Hypothesized
Stated, not yet evidenced here: electricity access allows productive uses (lighting, machinery, refrigeration, communications) that raise output per person. The catalog carries both series; no causal study is attached.
- Electricity system enables Human health and longevity Hypothesized
Stated, not yet evidenced here: electricity access allows refrigeration of vaccines and food, lighting in clinics, water pumping and a shift away from solid cooking fuels, all of which bear on mortality. The catalog carries both series; no study is attached.
Unknowns
- Data-center load growth is announced, not metered; the stored series cannot separate it from other demand.
- Interconnection queue withdrawal rates are not ingested, so the share of the planned pipeline that completes is not known from stored data.
Section 4 of 9
What has not changed?
AI narrative Modeled
Average retail electricity price at 14.48 cents per kilowatthour shows year-over-year change inside the 10th to 90th percentile band of past changes [obs:410784]. Capacity utilization at 76.8504 percent has change inside its historical band [obs:415537]. Population with access to electricity at 91.9264843853681 percent is within its band [obs:28425].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
| Series | Latest | Basis | Normal band (p10 to p90) | Historical range | Reading |
|---|---|---|---|---|---|
| US average retail electricity price Observed Average retail price; a demand squeeze shows up here first. | 14.48 cents per kilowatthour 2026-06-01, obs 410784 | year-over-year change 0.62 (+4.5%) 515 past changes | -0.096 to 0.77 | 3 (1976-07-01) to 14.48 (2026-06-01) since 1976-07-01; latest at the 100th percentile of levels | within its normal band |
| US capacity utilization Observed Industrial capacity utilisation. | 76.85 percent of capacity 2026-09-01, obs 415537 | year-over-year change 0.8043 (+1.1%) 704 past changes | -4.61 to 3.79 | 64.08 (2020-04-01) to 89.39 (1967-01-01) since 1967-01-01; latest at the 25th percentile of levels | within its normal band |
| Population with access to electricity Observed Share of people with electricity, world. | 91.93 percent of population 2024, obs 28425 | year-over-year change 0.323 (+0.4%) 25 past changes | 0.2347 to 1.13 | 73.2 (1998) to 91.93 (2024) since 1998; latest at the 100th percentile of levels | within its normal band |
Section 5 of 9
What absorbs the shock?
AI narrative Modeled
Evidenced buffers include nameplate capacity of 1310.444 million kilowatts above peak demand [obs:410776] and battery storage additions of 743.4 MW [obs:410640]. Financing cost is set by the federal funds rate at 3.75 percent [obs:410817]. Substitutes for power transformers are copper and aluminum. Imports, inventories and demand elasticity are not measured.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
US electric net summer capacity
1,310.44 million kilowatts · 2026-06-01 · yoy +4.7% · obs 410776
Nameplate capacity above peak demand is the first buffer; the stored series is the total, not the reserve margin.
743.4 MW · 2026-08-01 · yoy -29.5% · obs 410640
Battery additions shift energy within the day; they do not add annual energy.
Power transformer
US federal funds effective rate
3.75 percent · 2026-09-01 · yoy -11.1% · obs 410817
The policy rate sets the cost of capital for generation and grid projects.
Not measured
No stored series evidences this buffer.
Cross-border power imports are not ingested.
Not measured
No stored series evidences this buffer.
Transformer and switchgear inventories are not published as a series.
Not measured
No stored series evidences this buffer.
Price response of demand is not measured here.
Section 6 of 9
What response is underway?
AI narrative Modeled
The modeled 24-month capacity pipeline stands at 112974.47 MW with 21.2 percent year-over-year growth [obs:410691]. The modeled capacity response ratio is 0.968 [obs:410706]. Corporate grid investment capex growth is 30.02 percent [obs:409138]. Nonresidential construction spending rose 0.5 percent to 1308532 millions of dollars [obs:410855]. Planned retirements are 20 MW [obs:403974].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
| Group | Series | Latest | Year earlier | Change |
|---|---|---|---|---|
| Capacity pipeline | US future capacity pipeline, 24 months Modeled | 112,974.47 MW (2026-08-01, obs 410691) | 93,240.11 (2025-08-01) | 19,734.37 (+21.2%) |
| Capacity pipeline | US capacity response, electricity Modeled | 0.968 ratio (2026-08-01, obs 410706) | 0.8893 (2025-08-01) | 0.0787 (+8.8%) |
| Capacity pipeline | US planned generation retirements Observed | 20 MW (2072-08-01, obs 403974) | 1.2 (2065-12-01) | 18.8 (+1566.7%) |
| Corporate pulse | Corporate pulse: grid investment capex growth Derived | 30.02 percent change yoy (2026-06-30, obs 409138) | n/a | n/a |
| Capital spending | US nonresidential construction spending Observed | 1.31M millions of dollars, seasonally adjusted annual rate (2026-08-01, obs 410855) | 1.3M (2025-08-01) | 6,543 (+0.5%) |
The US pipeline and response measures are on the capacity page; bottleneck assessments are on bottlenecks.
Section 7 of 9
Scenarios
AI narrative Modeled
The 12-month base case forecast for US electricity net generation is 360015.27829501714 million kilowatthours with 80 percent interval 350484.2848454142 to 374291.7815749615; falsifier is first published value outside that interval [forecast:ae943369-165a-4a14-a64b-8fd82bb09397]. The 6-month base case is 334985.60400000017 million kilowatthours with interval 331672.44569794775 to 354002.5014654448 [forecast:85ecfac4-bd55-4314-95b1-80a7c92ef3f8].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
US electricity net generation
Base case
360,015.28 million kilowatthours
The point estimate: the model's central path.
Upside
374,291.78 million kilowatthours
Upper bound of the interval: what would need to go better than the central path.
Downside
350,484.28 million kilowatthours
Lower bound of the interval: what would need to go worse than the central path.
The model assigns 80 percent to the interval; the site assigns no other probability.
Falsifiers and assumptions
- If the first published value of us-electricity-net-generation (USA) for 2027-05 is below 350,484 million kilowatthours or above 374,292 million kilowatthours, the 80 percent interval is falsified.
- The model assigns a 10 percent probability to a value below 350,484 million kilowatthours and a 10 percent probability to a value above 374,292 million kilowatthours; interval misses should occur about one time in five over many forecasts, and a run of misses well above that rate falsifies the calibration.
- The point forecast is above the last observed value (354,691 million kilowatthours); a realised value at or below that level falsifies the direction call.
- Assumption: No structural break in us-electricity-net-generation between the data cutoff and the target period.
- Assumption: The source keeps publishing the series on the same definition; a rebasing or redefinition invalidates the comparison.
- Assumption: The average historical change per period continues.
US electricity net generation
Base case
334,985.6 million kilowatthours
The point estimate: the model's central path.
Upside
354,002.5 million kilowatthours
Upper bound of the interval: what would need to go better than the central path.
Downside
331,672.45 million kilowatthours
Lower bound of the interval: what would need to go worse than the central path.
The model assigns 80 percent to the interval; the site assigns no other probability.
Falsifiers and assumptions
- If the first published value of us-electricity-net-generation (USA) for 2026-11 is below 331,672 million kilowatthours or above 354,003 million kilowatthours, the 80 percent interval is falsified.
- The model assigns a 10 percent probability to a value below 331,672 million kilowatthours and a 10 percent probability to a value above 354,003 million kilowatthours; interval misses should occur about one time in five over many forecasts, and a run of misses well above that rate falsifies the calibration.
- The point forecast is below the last observed value (354,691 million kilowatthours); a realised value at or above that level falsifies the direction call.
- Assumption: No structural break in us-electricity-net-generation between the data cutoff and the target period.
- Assumption: The source keeps publishing the series on the same definition; a rebasing or redefinition invalidates the comparison.
- Assumption: The seasonal pattern of the last year repeats.
Section 8 of 9
What to watch
AI narrative Modeled
The 24-month capacity pipeline at 112974.47 MW shows 21.1651 percent growth above threshold 0, rising confirms easing [obs:410691]. Capacity additions at 2849.2 MW show -39.8725 percent growth, rising confirms easing [obs:410638]. Planned retirements at 20 MW show 1566.6667 percent growth, falling confirms easing [obs:403974]. Retail price at 14.48 cents shows 4.4733 percent growth above threshold 3, falling confirms easing [obs:410784].
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
On the easing side of the threshold · rising confirms easing
A growing 24-month pipeline means additions are being brought forward faster than they complete.
On the worsening side of the threshold · rising confirms easing
Completed additions rising year over year confirm the pipeline is converting.
On the worsening side of the threshold · falling confirms easing
Fewer planned retirements keeps existing capacity on the system while new capacity arrives.
On the worsening side of the threshold · falling confirms easing
Retail price growth above general inflation would signal the squeeze is reaching consumers.
Section 9 of 9
Forecast track record
AI narrative Modeled
For the 12-month horizon the champion model is drift with backtest MAE 9312.13262801899 over 36 cutoffs and 80 percent interval coverage 89 percent. For the 6-month horizon the champion is seasonal_naive with MAE 11292.10130952381 over 42 cutoffs and coverage 86 percent.
Written by grok-4.3 over the numbers in this snapshot only, validated against them (every figure and cited id had to be in the snapshot), run b0727bbe, prompt 1.0.0. It adds no facts.
| Forecast | Champion | Backtest MAE | Skill vs last value | 80% coverage | Live evaluations |
|---|---|---|---|---|---|
| us-electricity-net-generation:USA:12 | drift | 9,312.13 over 36 cutoffs | +14% | +89% | none yet |
| us-electricity-net-generation:USA:6 | seasonal_naive | 11,292.1 over 42 cutoffs | +71% | +86% | none yet |
The full scorecard is at /forecasts/scorecard.
Provenance
- Snapshot
- 49 · 2026-10-19 00:22 UTC · navigator_view_v1+navigator_topics_v1
- Gaps stated
- 0
- AI narrative
- grok-4.3, run b0727bbe-a424-4f26-9640-003c6198b915, prompt 1.0.0, 25 ids cited
- Earlier snapshots
- 41 (2026-10-12 00:11 UTC), 33 (2026-10-05 00:11 UTC), 25 (2026-09-28 00:12 UTC), 17 (2026-09-21 00:12 UTC), 9 (2026-09-14 00:14 UTC), 1 (2026-09-07 22:35 UTC)
Every value on this page is a stored observation (obs ids shown), a forecast record, a mechanism or a relationship row, or a percentile computed from those the same way the relevance engine computes its features. Series the config references but the site does not store are listed as gaps under their section, never filled in.