The heat is at a record
The variable this series opened with is the one the year turns on. Global surface temperature reached its highest value in the instrumental record in 2024, at 1.28 degrees above the mid-twentieth-century baseline, capping a two-decade climb; 2025 stepped back slightly to 1.19 degrees as a weak La Nina cooled the Pacific, but remained among the warmest years ever measured. The heat penalty of part 01, the output a module gives up simply for running hot, is a direct function of this temperature, and it now sits at its structural maximum: an average operating cell temperature near 53 degrees and a static loss of most of a tenth of nameplate output before the smaller, slower warming drift is even counted.
| Year | Anomaly °C | ENSO state | Global precip mm |
|---|---|---|---|
| 2021 | +0.85 | La Nina, moderate | 1007 |
| 2022 | +0.90 | La Nina, strong | 1008 |
| 2023 | +1.17 | El Nino, onset | 1022 |
| 2024 | +1.28 (record) | El Nino, weak | 1059 (record) |
| 2025 | +1.19 | La Nina, weak | 1019 |
NASA GISTEMP v4 temperature and GPCP precipitation, with the ENSO state from the NOAA Oceanic Nino Index. 2024 is the hottest and wettest year in the 2004-2025 record; the year-to-year swing between them is the variability of part 03, not a trend reversal.
Solar is scaling into it
The second half of the state of the climate is the electricity the world generates, because that is what decides how much the heat and the sky trends matter at scale. The transition crossed two thresholds in the current record. Total renewables reached 34 per cent of global generation and drew level with coal at 33.6 per cent. And solar, which was a rounding error in 2004, reached 8.7 per cent and overtook wind to become the largest single renewable source on the planet.
| Source | 2004 share | 2025 share |
|---|---|---|
| Coal | 40% | 33.6% |
| Renewables (total) | 18% | 34% |
| Solar | ~0% | 8.7% |
| Wind | 0.7% | 8.4% |
Ember Global Electricity Review; 2004 is a composite estimate (data_version composite_est), 2025 is Ember-native. Renewables total includes hydro. Solar overtook wind, and renewables drew level with coal, in the current record.
The two readings belong together. Every point of share that solar takes is a point of the world's electricity that now runs through the production climate this series maps: exposed to the heat penalty, the sky trends and the rest of the series' variables. With renewables at 34 per cent of global generation and solar the largest single source among them, solar's climate sensitivity has become a first-order question for grid planning as much as for the individual roof.
The nine readings
The series in one view. Each variable at its latest reading, and its direction of travel, with the full treatment a click away.
| Variable | Reading, 2026 | Direction |
|---|---|---|
| 01 · Heat penalty | 53°C mean cell temperature; 9.8% static loss (PERC); +0.34% added by year 25 of warming | Rising |
| 02 · Brightening and dimming | 26 of 40 markets brightening over two decades, 14 dimming; +0.4% on average | Mixed by region |
| 03 · The variable years | 2025 a weak La Nina after the 2023-24 El Nino; interannual swing 2.5% to 8% | Cyclic, no trend |
| 04 · Beam and diffuse | Diffuse share 19% (Atacama) to 63% (Reykjavik); sets where trackers pay | Structural |
| 05 · Rain, dust and soiling | 1.5% temperate to a 20% Gulf lower bound; drying markets lengthen the accumulation window | Stable, rising in dry belts |
| 06 · Smoke | A recurring seasonal factor in three regions; fire weather trending up where it reaches grids | Rising |
| 07 · Wind cooling | Refunds 3.6% to 6.9% of output; independent of temperature (r = -0.09) | Stable |
| 08 · Storms and extremes | Hail high-hazard in 2 markets, 5 exempt; hail-days rising in Sydney and Perth | Mixed, rising in parts |
| 09 · The compound trajectory | Value per kW $40 (Cairo) to $514 (Los Angeles); the tariff spread is 20 to 1 | Market-driven |
Each reading is the headline figure from that part of the series; the full data, method and buyer guidance sit behind each link. Directions describe the physical or market trend, not a forecast.
One argument the series has settled
Taken as a set, the nine readings settle one argument the series has made from the start. Of everything that moves the value of a solar kilowatt, the physical climate, the sum of parts 01 through 08, varies it by roughly a factor of two from the best market to the worst. The local price of power in part 09 varies it by a factor of twenty.
20 to 1
The local price of power varies the value of a solar kilowatt twenty-fold. The entire physical climate, parts 01 through 08 combined, varies it by about two.
The physical state is demanding and slowly worsening: record heat, rising fire weather, lengthening dry spells in the markets that soil hardest (part 05). The market side is where the speed is: solar's share of world electricity went from a rounding error to 8.7 per cent inside two decades. The heat and the storms decide which module and which mount, the sky trend and the variability decide which baseline to believe, and the tariff, more than anything physical, decides what the kilowatt is worth.
The practical summary of the whole series: sunlight is one input of nine. A solar system is a twenty-five-year instrument, the horizon this series models throughout, and on the series' own numbers the sunniest place is neither the safest nor the most valuable.
A synthesis, not a new measurement
This is a synthesis, not a new measurement. Every figure is the current reading of a pillar published in full elsewhere in the series, drawn from the same public records: NASA for temperature, irradiance, wind and the reanalysis behind the yield model; NOAA and partner agencies for the ocean and circulation drivers; Ember for the generation mix; and published retail tariffs for the demand side. The temperature and mix figures are the latest complete-year values in the record at the time of writing, mid-2026, on the 2024 and 2025 data. This report is designed to be refreshed as each year's record lands; the nine-variable structure is stable while the numbers update. Where a reading rests on a modelled or literature-derived figure rather than a direct measurement, that limit is stated in the underlying part. Everything here is the record as of mid-2026.
A July 2026 Caltech-Google large-eddy simulation study adds new mechanism evidence about how tropical Pacific low clouds may amplify warming, assessed separately in Pacific Low-Cloud Feedback Under Warming and High CO2. Its regional model responses do not replace any observed reading in this annual state.