When France gets cold, its neighbours feel it at the border. Each degree of French winter cold takes about 1 GW off French net exports and goes with day-ahead prices €5 ($5.6)/MWh to €7/MWh higher in France, Belgium and Germany. Roughly half of the French price response, a third of the Belgian and a quarter of the German is French temperature on its own rather than the neighbours’ weather. Eleven winters of data can also say why the loss sits in the hours solar cannot fill, why the nuclear fleet rather than the thermometer decides whether France keeps exporting, and why a cold spell can leave a solar plant better off.

01. One cold spell, and the pattern of eleven winters

The winter of 2025/26 showed the whole mechanism in four weeks. Between 15 and 21 December 2025, with the French effective temperature, a population-weighted average of the day’s temperature and the two days before it, at 9.8 C, France exported a net 16.3 GW on average to its six neighbours. Between 29 December and 11 January, at 1.9 C, it exported 5.7 GW. On 6 January, at −1.5 C and with French demand averaging 82 GW over the day, France imported a net 3.7 GW. Nuclear output was 52.4 GW in the mild week and 53.3 GW in the cold spell, wind averaged 7.2 and 7.5 GW, and French solar averaged no more than 2.3 GW on its best day of the cold spell. The swing in exports was therefore driven by demand, which rose from 56 GW to 72 GW between the two windows. Lower exports covered about 10.6 GW of that 16 GW rise. The remaining 5 GW or so was met mostly by gas, with smaller gains in nuclear, hydro, solar and wind.

8 GW to 9 GW between a mild day and a cold one

That cold spell is close to the average pattern of eleven winters, counting November to March as winter. On the mildest tenth of winter days, above 11.6 C, France exports 9.0 GW. On the coldest tenth, below 3.2 C, it exports close to zero. Measured within each winter rather than across all eleven, the swing is 8.3 GW, so the loss between a mild day and a cold one is 8 to 9 GW whichever way it is measured. Germany loses 2.1 GW of French supply between the two, on physical flows that include loop flows, Belgium 1.9 GW, Switzerland 1.7 GW, Britain 1.3 GW, Spain 1.1 GW and Italy 1.0 GW, and on the German, Belgian and Spanish borders the flow reverses on average. On cold days the Belgian, German and Swiss borders run near their highest winter import levels in 5 to 6% of hours, against under 1% on mild days. In a cold spell the question shifts from how much France can export to how much it can take in.

02. Where the degree goes

Both the December spell and the eleven-winter average put the swing down to demand, so the share of the extra demand that is met at home decides how much of the loss crosses the border. French demand starts to respond to cold below 15.5 C. Within a winter, once month, weekday and holidays are allowed for, it rises by 1.8 GW for every degree colder. That response has not weakened, at 1.7 GW per degree in 2015/16 and 1.9 GW in 2025/26. Between 15.5 C and freezing, demand climbs by about 30 GW. Below 2 C the average winter day turns France from exporter to importer.

Of the 1.8 GW a degree adds within a winter, nuclear output is 0.43 GW higher, gas 0.36 GW, hydro 0.20 GW, coal and oil 0.11 GW, and net exports 1.04 GW lower. Wind falls by 0.35 GW, as cold French days tend to be still ones, so that loss has to be covered as well, and the five supply terms add up to the 1.8 GW rise in demand plus the 0.35 GW lost from wind. Solar shows no response to temperature within a winter. At 1.04 GW per degree, a winter 1 C colder than the 7.8 C average would remove about 3.8 TWh from net exports.

The neighbours are cold on the same days

The obvious objection is that the neighbours are cold on the same days, so the flows reflect their weather as much as France’s. Holding their temperatures and the French and German wind and solar forecasts fixed gives a larger export response, not a smaller one. The reason is that the neighbours’ own cold pulls French power across the border on the same days, which partly hides the loss in the raw figures. With those controls in place, French demand rises 1.48 GW per degree of French cold rather than 1.8 GW, nuclear output no longer moves with the French thermometer, and 1.29 GW of that 1.48 GW is met by lower exports. The 0.43 GW of nuclear uplift per degree in the raw figures therefore goes with the neighbours’ pull on French supply rather than with French cold itself. The 1.04 GW is the figure for a winter in which the neighbours are cold at the same time, as they usually are; the 1.29 GW is the French temperature effect alone, and it falls almost entirely on the borders. The split between French and neighbouring cold is approximate, because the two move together.

Each border loses between 0.16 and 0.28 GW per degree

On that basis, Belgium loses 0.28 GW per degree of French cold, Germany 0.26 GW, Switzerland 0.21 GW, Italy and Britain 0.19 GW and Spain 0.16 GW, and the six borders sum to the 1.29 GW. Belgian and French effective temperatures move almost in step, so Belgium’s coldest days and the loss of French supply arrive together.

03. The fleet decides how much France can spare

The per-degree figures describe an average winter. Whether France keeps exporting through a particular cold spell depends on what the nuclear fleet is producing when the cold arrives, and the coldest fortnight of each winter shows what France could still spare. In December 2022, at 2.2 C with nuclear output at 40 GW, France imported 8.0 GW. In January 2025, at 3.0 C with nuclear output at 55 GW, it exported 7.2 GW. The coldest fortnight of 2025/26, straddling the turn of the year, was colder than both, at 1.3 C, and France exported 5.9 GW with 54 GW of nuclear output. Cold of much the same depth has produced a 15 GW range in French net exports. Even with nuclear output at 55 GW, the deepest cold of the eleven winters, 0.7 C in January 2017, pulled France into import, though only by 2.0 GW.

The seasonal balance tells the same story. It was −4 TWh in 2021/22 and close to zero in 2022/23, the two winters with nuclear output averaging 42 GW or less. It reached 44 TWh in 2025/26, a winter that was milder than average and had 49 GW of average nuclear output. France was a net importer on 58% of winter days in 2021/22 and on 1% in 2025/26.

04. Solar in a French cold spell

French solar output has more than quadrupled since 2015, so the natural question is whether it can fill any of that winter gap. Its share of French winter generation has risen from under 1% to 3.2%, but December and January output has been 26% to 37% of the summer before it in every year since 2015.

On the coldest tenth of winter days French solar averaged 1.0 GW against 1.3 GW on the mildest, and German solar, on the day-ahead forecast, 2.0 GW against 3.7 GW. The coldest days cluster in December and January, when days are shortest. Within those two months a cold day has more French solar than a mild one, so the lower output on cold days is a matter of the calendar, not of the cold. In the last two winters French solar covered 2.6% of demand on the coldest days. Doubling capacity would take that to about 5%, against a 23 GW gap in demand between the mildest and coldest tenth of winter days.

The loss sits in the hours solar cannot reach

Estimated hour by hour, the export loss peaks at 1.18 GW per degree at 08:00 and at 1.10 GW at 19:00. Both are hours in which winter solar output is close to zero. French demand itself responds least to cold in mid-afternoon, and the export loss follows it, with a trough of 0.93 GW at 15:00.

05. Price formation on a cold French day

On the coldest tenth of winter days the French day-ahead price averaged €113/MWh, against €61/MWh on the mildest. The neighbours moved with it. Belgium averaged €111/MWh against €66/MWh and Germany €101/MWh against €66/MWh, and the French premium over Germany flipped from −€5/MWh on mild days to 12 €/MWh on cold days. Gaps measured across all eleven winters depend on which winters the cold days happen to fall in, and without 2022/23 the French gap is €38/MWh rather than €52/MWh, so the per-degree response, estimated within each month of each winter, is the better guide.

Per degree of French cold, estimated within each month of each winter, the French baseload price rises €6.7/MWh, the Belgian €6.0/MWh, the German €5.3/MWh, the Dutch and Swiss €4.7/MWh, the northern Italian €3.2/MWh and the Spanish €2.8/MWh. In every winter since 2015 the French price rise per degree has been between 4% and 9% of that winter’s average price.

Holding the neighbours’ temperatures and the wind and solar forecasts fixed, as in the export estimate, the French response falls to €3.7/MWh, the Belgian to €2.0/MWh and the German to €1.4/MWh. On these estimates, roughly half of the French price response, a third of the Belgian and a quarter of the German is French cold on its own, and the rest comes with the neighbours’ own cold and the low wind that arrive on the same days.

Within the day, the French price response peaks at €8.0/MWh per degree at 08:00 and €8.4/MWh at 18:00. The Belgian and German curves follow the same shape. The night hours respond least, and midday responds less than either peak.

Capture rates rise on cold days; revenue depends on geography and on what has been sold

That shape has a consequence for solar. The capture rate is the average price solar receives divided by the average price for all hours. It rises in all seven zones on the coldest tenth of winter days compared with the mildest, and reaches parity or above everywhere except Spain. France moves from €0.92 to €1.03, Belgium from €0.93 to €1.03, Germany from €0.96 to €1.05, the Netherlands from €0.91 to €1.04, Switzerland and northern Italy from €0.98 to €1.03, and Spain from €0.82 to €0.97. With little solar on a cold day there is no midday surplus to push the price down, and the first solar hours fall just after the morning peak, when prices are still high, so the price solar receives rises faster than the average and ends up above it. The French solar capture price doubles, from €58 to €118, and the Dutch almost doubles, from €73 to €142.

The price data cover fewer winters than the flow data, and on their cold days solar output is 28% lower than on their mild days in France, 42% lower in Belgium and 52% lower in Germany, a slightly larger drop than in the eleven-winter comparison. Taking average capture price multiplied by average output, French solar earned 47% more per day on the coldest days than on the mildest, because the price gain outran the output loss. Spanish solar earned 13% more and Belgian solar 8% more. German solar earned 15% less, even with baseload prices more than 50% higher, because the output loss was larger than the price gain. Whether a portfolio gains or loses depends as much on how much winter output it has already sold, and in what hourly profile, as on where the plant sits. A producer who has sold a fixed winter volume forward is short that output on exactly the days when buying it back costs most, and the buyer of a pay-as-produced contract holds the opposite position. On these data a cold spell is not a capture-rate problem for winter solar but a volume one.

The storage spread moves from the solar cycle to the overnight cycle

A battery that charges on cheap midday solar and discharges into the evening trades the midday-to-evening spread. That spread narrows on cold days, from €27/MWh to €20/MWh in France and from €34/MWh to €17/MWh in Germany, because the midday price rises with everything else once there is no solar to push it down. The overnight-to-morning spread moves the other way, from €30/MWh to 43 €/MWh in France and from €28/MWh to €41/MWh in Germany between mild and cold days, because the night hours respond least and the morning peak responds far more. A cold spell therefore moves the day-ahead spread a battery can trade in winter from the solar cycle to the overnight cycle rather than removing it, and a battery able to cycle twice a day can trade both.

06. Conclusion

Eleven winters of data make French winter cold a measurable quantity for the markets around it. Each degree takes about a gigawatt off French exports, the six neighbours share that loss, and the French price response has stayed between 4% and 9% of the winter price level in every winter since 2015, including the crisis winters. A model of Belgian or German winter prices that carries local temperature but not French temperature leaves out roughly a quarter to a third of the price response to a French cold day.

For solar, winter output is set by the calendar rather than by the cold, and cold-day prices rise, so the exposure sits in how winter output has been sold rather than in the capture rate. For storage, a cold spell moves the earning window from the midday-to-evening cycle to the overnight-to-morning cycle.

Two forces pull in opposite directions over the coming winters. A fuller French fleet leaves more to export through a cold fortnight, as the last two winters show. Against that, the French demand response per degree is no lower than it was eleven winters ago, at 1.9 GW in 2025/26 against 1.7 GW in 2015/16. Solar growth does little for the peaks in which the loss and the price response are largest. The balance between fleet availability and demand decides whether the next cold spell looks like the turn of 2025/26, with France still exporting 5.9 GW through its coldest fortnight, or like December 2022, with France importing 8.0 GW.

How WSP’s Electricity Market Outlook can help

WSP’s Electricity Market Outlook (EMO) is designed to analyse these market dynamics. The underlying PRIMES model has supported European Commission energy-policy analysis for more than two decades and simulates all major European electricity markets simultaneously through to 2050. Cross-border flows are derived using a replication of the EUPHEMIA algorithm used in Europe’s single day-ahead coupling. Model outputs include hourly wholesale prices, capture rates, negative-price frequency and depth, curtailment exposure, interconnector utilisation and BESS profitability projections at both country and asset level. These outputs provide the quantitative basis for winter risk assessment, PPA and CfD bid analysis and long-term storage revenue modelling.

Author: Safa Sen, Market Engagement Lead For CWE at WSP.

WSP Group unites engineering, advisory and science-based expertise to shape communities to advance humanity. ​ ​From local beginnings to a globe-spanning presence today, it operates in over 50 countries and provides solutions and delivers innovative projects across sectors: Transport & Infrastructure, Property & Buildings, Earth & Environment, Water, Power & Energy and Mining & Metals.