HUMANITYVECTOR

World · snapshot 2026-09-07 22:36 UTC · data to 2025

CO2 and the energy transition

Is atmospheric CO2 still accelerating, and is the capacity being built enough to change that?

Atmospheric CO2, global annual mean was 425.62 ppm in 2025, up 0.7 percent over one year, a move at the 96th percentile of its one-year changes.

DerivedModeledPrimary metric page

Section 1 of 9

What changed?

Observed and derived series only. Changes are latest minus the value one, five and ten years earlier.

AI narrative Modeled

Atmospheric CO2 global annual mean reached 425.62 ppm in 2025 [obs:264263], up 2.83 ppm or 0.7 percent from 422.79 ppm in 2024 at the 96th percentile of one-year changes [obs:264262]. Five-year change was 13.18 ppm or 3.2 percent [obs:264258] and ten-year change 25.97 ppm or 6.5 percent [obs:264253]. Global surface temperature anomaly was 1.19 degrees Celsius in 2025 [obs:264216]. CO2 emissions per person stood at 4.69378456136626 tonnes in 2024 [obs:15882].

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 24aa0684, prompt 1.0.0. It adds no facts.

Observed

Global mean CO2, NOAA.

425.62 ppm

2025

47 points, 1979 to 2025, 336.85 to 425.62 ppm336.85425.6219792025
SpanFromChangePercentile
1 period422.79 (2024)2.83 (+0.7%)96th
1 year422.79 (2024)2.83 (+0.7%)96th of 45
5 years412.44 (2020)13.18 (+3.2%)100th of 41
10 years399.65 (2015)25.97 (+6.5%)100th of 36
World · 47 observations since 1979 · obs 264263
Observed

Global temperature anomaly.

1.19 degrees Celsius vs 1951-1980

2025

60 points, 1966 to 2025, -0.1 to 1.28 degrees Celsius vs 1951-1980-0.11.2819662025
SpanFromChangePercentile
1 period1.28 (2024)-0.09 (-7.0%)57th
1 year1.28 (2024)-0.09 (-7.0%)57th of 144
5 years1.01 (2020)0.18 (+17.8%)72nd of 140
10 years0.9 (2015)0.29 (+32.2%)84th of 135
World · 146 observations since 1880 · obs 264216
Supporting series

CO2 emissions per person

Observed

CO2 emissions per person.

4.69 tonnes CO2 per person

2024

55 points, 1970 to 2024, 3.94 to 4.81 tonnes CO2 per person3.944.8119702024
SpanFromChangePercentile
1 period4.68 (2023)0.0162 (+0.3%)17th
1 year4.68 (2023)0.0162 (+0.3%)17th of 53
5 years4.7 (2019)-0.0074 (-0.2%)12th of 49
10 years4.78 (2014)-0.0835 (-1.7%)18th of 44
World · 55 observations since 1970 · obs 15882
  • co2-mauna-loa-monthly-mean (WLD): no observations, no current observations for WLD

Section 2 of 9

How unusual is it?

The primary series against its own stored history (features_v1 rarity and surprise).

AI narrative Modeled

Latest value sits at the 100th percentile of series history, an all-time high [obs:264262]. One-year change is at the 96th percentile with robust z of 1.5007 and one period since a larger move [obs:264262]. Prior maximum was 422.79 ppm in 2024 [obs:264262] and prior minimum 336.85 ppm in 1979 [obs:264217].

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 24aa0684, prompt 1.0.0. It adds no facts.

Derived
Latest value percentile
100th of 46 values
Latest change percentile
96th of 45 changes
Robust z of latest change
1.50
Largest move
1 period since a larger move
Record level
all-time high in the stored series
Previous comparable move
2024: 3.44 ppm (obs 264262, 264261)
Prior maximum
422.79 (2024, obs 264262)
Prior minimum
336.85 (1979, obs 264217)
Derived from the stored series; percentiles are rank percentiles with ties split.

Section 3 of 9

Why might it be happening?

Documented mechanisms, statistical model drivers, hypotheses and unknowns are kept apart.

AI narrative Modeled

Documented mechanisms are not stored. Statistical model drivers are not stored. Hypotheses include electricity system enabling economic output and human health and longevity [rel:e401fb30-2d25-4455-adf6-5b1a95069d34] [rel:f913d900-026e-4232-b9aa-a617096de9f8]. Unknowns: world fossil generation and capacity additions not ingested; lag from emissions to concentration not modeled.

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 24aa0684, prompt 1.0.0. It adds no facts.

Reviewed

Documented mechanisms

No mechanism in the mechanisms table is linked to this topic yet.

Statistical

Model-supported drivers

Modeled

No active forecast for this topic records driver contributions (the champion may be a baseline with no drivers).

Candidate or hypothesized relationships

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.

Not settled by stored data

Unknowns

  • World fossil generation and world capacity additions are not ingested; the response section shows US series only.
  • The lag from emissions to concentration to temperature is documented physics, but the site does not store a model of it.

Section 4 of 9

What has not changed?

Fundamentals judged against the 10th to 90th percentile band of their own year-over-year changes (or of their levels when the history is short).

AI narrative Modeled

CO2 emissions per person year-over-year change of 0.016185 tonnes lies inside the 10th to 90th percentile band of -0.1017 to 0.128359 over 53 changes since 1970, with level band 4.007067 to 4.700061 [obs:15882]. US primary energy consumption change of -1.239 trillion Btu lies inside -295.7488 to 362.9531 band over 628 changes [obs:248499].

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 24aa0684, prompt 1.0.0. It adds no facts.

SeriesLatestBasisNormal band (p10 to p90)Historical rangeReading
CO2 emissions per person Observed
Per-person emissions have been roughly flat for a decade.
4.69 tonnes CO2 per person
2024, obs 15882
year-over-year change 0.0162 (+0.3%)
53 past changes
-0.1017 to 0.12843.94 (1994) to 4.81 (2013)
since 1970; latest at the 87th percentile of levels
within its normal band
US primary energy consumption Observed
US primary energy consumption.
7,414.83 trillion Btu
2026-05-01, obs 248499
year-over-year change -1.24 (-0.0%)
628 past changes
-295.75 to 362.955,241.8 (1975-06-01) to 9,521.89 (2025-01-01)
since 1973-01-01; latest at the 49th percentile of levels
within its normal band

Section 5 of 9

What absorbs the shock?

Each buffer names the series or graph entity that evidences it, or says it is not measured.

AI narrative Modeled

Reserves not measured. Inventory not measured. Demand elasticity not measured. Substitution shown by US solar additions of 1863.4 MW [obs:402440], wind additions of 0 MW [obs:403808] and battery additions of 1107.7 MW [obs:398336]. Lithium top-3 country concentration is 72.4 percent [obs:404808].

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 24aa0684, prompt 1.0.0. It adds no facts.

Observed

1,863.4 MW · 2026-07-01 · yoy +21.9% · obs 402440

Solar additions displace fossil generation, US.

Observed

0 MW · 2026-07-01 · yoy -100.0% · obs 403808

Wind additions, US.

Observed

1,107.7 MW · 2026-07-01 · yoy -33.9% · obs 398336

Battery additions, US.

Reserves

Not measured

No stored series evidences this buffer.

lithium-reserves is referenced but not available (see gaps). Lithium reserves for batteries.

Derived

72.4 percent of world production · 2025 · yoy -7.1% · obs 404808

Three-country share of lithium output.

Inventory

Not measured

No stored series evidences this buffer.

Fuel stocks are not ingested.

Demand elasticity

Not measured

No stored series evidences this buffer.

  • lithium-reserves (WLD): not published, metric is registered but not published

Section 6 of 9

What response is underway?

Capacity pipeline, corporate pulses, capital spending and production series, with their year-over-year change.

AI narrative Modeled

US future capacity pipeline stands at 109333.015 MW over 24 months, up 14.5 percent year over year [obs:409101]. Lithium mine production reached 290000 metric tons, up 30.6 percent [obs:263799]. Nickel mine production reached 3900000 metric tons, up 5.1 percent [obs:263830]. Corporate grid investment capex growth was 35.21 percent [obs:404027].

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 24aa0684, prompt 1.0.0. It adds no facts.

GroupSeriesLatestYear earlierChange
Capacity pipelineUS future capacity pipeline, 24 months Modeled109,333.02 MW (2026-07-01, obs 409101)95,452.66 (2025-07-01)13,880.36 (+14.5%)
ProductionLithium mine production Observed290,000 metric tons (2025, obs 263799)222,000 (2024)68,000 (+30.6%)
ProductionNickel mine production Observed3.9M metric tons (2025, obs 263830)3.71M (2024)190,000 (+5.1%)
Corporate pulseCorporate pulse: grid investment capex growth Derived35.21 percent change yoy (2025-12-31, obs 404027)n/an/a

The US pipeline and response measures are on the capacity page; bottleneck assessments are on bottlenecks.

Section 7 of 9

Scenarios

Base is the champion model's point estimate; upside and downside are the bounds of its 80 percent interval. No probability beyond the interval's own is assigned.

AI narrative Modeled

Base case for 2026 is 428.32781861213255 ppm with 80 percent interval 428.1238040127703 to 429.33460664210685, model holt [forecast:64f8c21f-8356-4fb0-9c40-a383cf0c41be]. Base case for 2027 is 431.0471947333191 ppm with interval 430.81780665910446 to 432.4847769528827 [forecast:80ef0dbc-f64f-4c48-85d7-800082e98b2f]. Base case for 2030 is 439.2053230968788 ppm with interval 438.8281030744282 to 441.8942115791888 [forecast:8b066c18-d8e5-4268-9b03-fe010b8fc9ee].

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 24aa0684, prompt 1.0.0. It adds no facts.

1 year horizon · target 2026

Atmospheric CO2, global annual mean

Modeled
47 observed points to 2025, then 1 forecast with 80 percent intervals336.9383.1429.3197919851991199720032009201520212026
Solid: observed (ppm). Dashed and shaded: model point estimates and 80 percent intervals from empirical residual quantiles. Filled dot: the record on this page. Green or red dot: the realised value, inside or outside its interval.

Base case

428.33 ppm

The point estimate: the model's central path.

Upside

429.33 ppm

Upper bound of the interval: what would need to go better than the central path.

Downside

428.12 ppm

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 co2-global-annual-mean (WLD) for 2026 is below 428.1 ppm or above 429.3 ppm, the 80 percent interval is falsified.
  • The model assigns a 10 percent probability to a value below 428.1 ppm and a 10 percent probability to a value above 429.3 ppm; 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 (425.6 ppm); a realised value at or below that level falsifies the direction call.
  • Assumption: No structural break in co2-global-annual-mean 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 smoothed level and trend at the cutoff persist over the horizon.
Record 64f8c21f · holt · data cutoff 2026-09-07 22:16 UTC · created 2026-09-07 22:16 UTC · WLD
2 year horizon · target 2027

Atmospheric CO2, global annual mean

Modeled
47 observed points to 2025, then 1 forecast with 80 percent intervals336.9384.7432.5197919851991199720032009201520212027
Solid: observed (ppm). Dashed and shaded: model point estimates and 80 percent intervals from empirical residual quantiles. Filled dot: the record on this page. Green or red dot: the realised value, inside or outside its interval.

Base case

431.05 ppm

The point estimate: the model's central path.

Upside

432.48 ppm

Upper bound of the interval: what would need to go better than the central path.

Downside

430.82 ppm

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 co2-global-annual-mean (WLD) for 2027 is below 430.8 ppm or above 432.5 ppm, the 80 percent interval is falsified.
  • The model assigns a 10 percent probability to a value below 430.8 ppm and a 10 percent probability to a value above 432.5 ppm; 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 (425.6 ppm); a realised value at or below that level falsifies the direction call.
  • Assumption: No structural break in co2-global-annual-mean 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 smoothed level and trend at the cutoff persist over the horizon.
Record 80ef0dbc · holt · data cutoff 2026-09-07 22:16 UTC · created 2026-09-07 22:16 UTC · WLD
5 year horizon · target 2030

Atmospheric CO2, global annual mean

Modeled
47 observed points to 2025, then 1 forecast with 80 percent intervals336.9389.4441.9197919851991199720032009201520212030
Solid: observed (ppm). Dashed and shaded: model point estimates and 80 percent intervals from empirical residual quantiles. Filled dot: the record on this page. Green or red dot: the realised value, inside or outside its interval.

Base case

439.21 ppm

The point estimate: the model's central path.

Upside

441.89 ppm

Upper bound of the interval: what would need to go better than the central path.

Downside

438.83 ppm

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 co2-global-annual-mean (WLD) for 2030 is below 438.8 ppm or above 441.9 ppm, the 80 percent interval is falsified.
  • The model assigns a 10 percent probability to a value below 438.8 ppm and a 10 percent probability to a value above 441.9 ppm; 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 (425.6 ppm); a realised value at or below that level falsifies the direction call.
  • Assumption: No structural break in co2-global-annual-mean 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 smoothed level and trend at the cutoff persist over the horizon.
Record 8b066c18 · holt · data cutoff 2026-09-07 22:16 UTC · created 2026-09-07 22:16 UTC · WLD

Section 8 of 9

What to watch

Each indicator is a stored series with a numeric threshold; the reading says which side of it the latest value sits on.

AI narrative Modeled

Atmospheric CO2 at 425.62 ppm shows 0.6694 percent year-over-year against 0.6 threshold, status worsening [obs:264263]. Temperature anomaly at 1.19 degrees Celsius against 1.5 threshold, status easing [obs:264216]. US solar additions at 1863.4 MW show 21.8546 percent year-over-year, status easing [obs:402440]. Lithium production at 290000 metric tons shows 30.6306 percent year-over-year, status easing [obs:263799].

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 24aa0684, prompt 1.0.0. It adds no facts.

Observed
Year over year: +0.7% (2025)Threshold: +0.6%

On the worsening side of the threshold · falling confirms easing

Annual growth below about 0.6 percent (roughly 2.5 ppm) would be slower than the last decade.

obs 264263, 264262 · World
Observed
Level: 1.19 degrees Celsius vs 1951-1980 (2025)Threshold: 1.5 degrees Celsius vs 1951-1980

On the easing side of the threshold · falling confirms easing

The anomaly against the 1.5 degree reference used in climate agreements.

obs 264216 · World
Observed
Year over year: +21.9% (2026-07-01)Threshold: 0%

On the easing side of the threshold · rising confirms easing

Solar additions growing year over year is the largest substitution channel in the stored data.

obs 402440, 402428 · United States
Observed
Year over year: +30.6% (2025)Threshold: +5%

On the easing side of the threshold · rising confirms easing

Storage depends on lithium output growing faster than 5 percent a year.

obs 263799, 263789 · World

Section 9 of 9

Forecast track record

Backtests are rolling-origin, out of sample; live evaluations compare published forecasts with the realised values.

AI narrative Modeled

For co2-global-annual-mean:WLD:1 champion holt backtest MAE is 0.3138980665319373 over 16 cutoffs with 80 percent interval coverage 75 percent and skill versus last value 87 percent. For horizon 2 MAE is 0.563228383256156 over 15 cutoffs with coverage 67 percent and skill 88 percent. For horizon 5 MAE is 1.4068605940130585 over 12 cutoffs with coverage 42 percent and skill 88 percent. No live evaluations yet.

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 24aa0684, prompt 1.0.0. It adds no facts.

ForecastChampionBacktest MAESkill vs last value80% coverageLive evaluations
co2-global-annual-mean:WLD:1holt0.3139 over 16 cutoffs+87%+75%none yet
co2-global-annual-mean:WLD:2holt0.5632 over 15 cutoffs+88%+67%none yet
co2-global-annual-mean:WLD:5holt1.41 over 12 cutoffs+88%+42%none yet

The full scorecard is at /forecasts/scorecard.

Provenance

Snapshot
7 · 2026-09-07 22:36 UTC · navigator_view_v1+navigator_topics_v1
Gaps stated
2
AI narrative
grok-4.3, run 24aa0684-c8c9-42a8-9f57-4a2f879795e8, prompt 1.0.0, 21 ids cited
Earlier snapshots
none

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.