Day-ahead markets across the EU-27 cleared 1,223 hours below zero in Q1 2026. That is more than double the 593 hours recorded in Q1 2025 and over ten times the recent low of 119 hours in Q1 2022. The aggregate masks divergent national stories. Spain alone accounted for 347 of those hours, while Finland and Sweden returned to zero after large prior-year surges. In April 2026, Germany cleared 123 of 720 hours below zero, and on the same days, day-ahead generation forecasts and actual delivered output diverged most sharply.
Q1 negative-price hours, 2019 to 2026
The Q1 series across the eight years from 2019 to 2026 shows where the structural pressure has built and where it has stabilised.
Q1 negative-price hours, day-ahead, by bidding zone
Image: ENTSO-E Transparency Platform
Spain is the most striking mover in the dataset. Before 2023, Spain had never recorded a Q1 negative-price hour. The number rose to 21 in 2023, climbed to 168 in 2024, eased back to 73 in 2025, and then jumped to 347 in Q1 2026, equivalent to 16 percent of all Q1 trading hours. Portugal followed the same Iberian pattern, reaching 294 hours in Q1 2026 from a 2023 baseline of 8. Greece moved from zero in Q1 2025 to 138 in Q1 2026, the largest single year-on-year increase in the table.
Poland began recording Q1 negative hours only in 2025, with 56 hours that year and 52 in 2026. Germany has settled into a 40 to 50 hour Q1 range across 2024, 2025 and 2026, well below its earlier peak of 131 hours in Q1 2019. Finland and Sweden’s SE2 zone each surged in 2024 and 2025, peaking at 85 and 41 hours respectively, before returning to zero in Q1 2026. That pattern reflects hydro storage and heating demand rather than solar oversupply, and it does not move in step with the Iberian trajectory. Italy has yet to record a single Q1 negative-price hour across the full eight-year series.
Q1 is also the mildest quarter of the year for solar-driven oversupply. Irradiance peaks in Q2 and Q3, and the Iberian and Greek patterns visible in Q1 extend further into the year. The Q1 numbers are therefore a floor on annual exposure rather than a representative quarter.
Depth of negative pricing, full-year 2025
Hour counts describe how often prices clear below zero. They do not describe how far below zero. Breaking down full-year 2025 negative-price hours into depth bands across the six core PPA markets shows that markets with comparable hour counts can produce very different price events.
Full-year 2025 negative-price hours by depth band, with mean and minimum
Image: ENTSO-E Transparency Platform
Germany recorded 576 negative hours in 2025 with a mean clearing price of −10.89 euros per MWh and a minimum of −250.32 euros per MWh. Thirteen hours fell below −100 euros per MWh. Spain logged 556 hours in total, but with a mean of just −2.10 euros per MWh and a minimum of −15.00 euros per MWh. No Spanish hour fell below −25 euros per MWh in the full year.
Poland is the most volatile of the six markets. Its 310 negative hours had a mean of −15.75 euros per MWh, the deepest of the group, and a minimum of −132.95 euros per MWh. Mean depth in Poland tripled between 2023 and 2025, rising from −4.94 euros per MWh to −15.75 euros per MWh. The Nordic markets sit at the opposite end of the distribution. Finland’s 465 negative hours averaged −1.25 euros per MWh with no hour falling below −25. Sweden’s SE2 zone showed a similar shallow profile across 681 hours. Greece logged 115 hours with a mean of −3.40 euros per MWh and a minimum of exactly −50.00 euros per MWh.
The same hour-count number can therefore describe very different financial events. A 556-hour Spanish year and a 310-hour Polish year are not comparable on price impact. Poland’s mean is roughly seven and a half times deeper than Spain’s, and its tail extends nine times further from zero.
Germany, April 2026: forecast error, curtailment, or both?
Negative prices in Germany are often framed as a straightforward renewables story. More wind and solar in a system with limited flexibility produces more zero-marginal-cost hours. April 2026 day-ahead data tells a more specific story. The hours in which prices cleared below zero are also the hours in which day-ahead generation forecasts and actual delivered output diverged most sharply.
Germany day-ahead market, April 2026. Selected metrics.
Image: ENTSO-E Transparency Platform
In April, 123 of 720 hours, or 17.1 percent of the month, cleared at a negative day-ahead price. Of those 123 hours, 105, or 85 percent, fell within the 10:00 to 16:00 CEST window, the strongest solar production block of the day. The mean clearing price across the 123 negative hours was −36.0 euros per MWh. The deepest single hour cleared at −413.8 euros per MWh.
That concentration shows up in the share of monthly generation delivered into negative prices. Solar produced 9,754 GWh in April, of which 4,559 GWh, or 46.7 percent, was delivered during negative-price hours. Wind produced 9,650 GWh, of which 1,536 GWh, or 15.9 percent, was delivered during negative-price hours. Almost half of the month’s solar output therefore monetised at a price below zero.
The forecast-actual gap was substantial in absolute terms. The combined absolute deviation between day-ahead forecast and actual generation reached 1,838 GWh in April, equivalent to 9.5 percent of total wind and solar output for the month. The relative gap was 6.2 percent for solar and 12.8 percent for wind, making the wind gap 2.07 times larger than the solar gap on a like-for-like basis.
The more revealing finding is that this gap is asymmetric across the price curve. During negative-price hours, actual generation fell below day-ahead forecast in 80 percent of solar hours, 98 of 123, and 83 percent of wind hours, 102 of 123. During positive-price hours, the same shortfall share dropped to 27 percent for solar and 49 percent for wind. Solar fell short of forecast nearly three times more often when prices were negative than when they were positive. Cumulative shortfall during negative-price hours alone reached 240 GWh for solar and 464 GWh for wind.
Two mechanisms produce that same shortfall, and the dataset cannot separate them. The first is genuine forecast error from weather diverging from the model used by the TSO and market participants ahead of gate closure. The second is curtailment, where turbines and inverters are throttled by the TSO or by market signals when prices crash. Both leave an identical signature in the data. Forecasted megawatts that never appear as delivered megawatts. Distinguishing between them requires plant-level dispatch data and curtailment instructions that are not published in this dataset.
The most extreme single hour of the month was 5 April at 14:00 CEST. Wind output was forecast at 39,156 MW; actual generation came in at 9,881 MW. The deviation of 29,275 MW coincided with a day-ahead price of −98.71 euros per MWh in that hour. April 5 was also the worst single day of the month, with a cumulative wind forecast-actual gap of 232 GWh.
Across one month and one bidding zone, the largest forecast-actual gaps align with the moments the market had already priced for abundance. The pattern is consistent rather than random.
How Ricardo’s Electricity Market Outlook can help
Do you need to anticipate when negative prices will hit your bidding zone, and how deep they will run? To know what share of your asset’s generation will land in negative-price hours? To bid a CfD that does not pay for curtailment?
Ricardo’s Electricity Market Outlook (EMO) is built to answer them. The underlying model, PRIMES-IEM, sits behind two decades of European Commission policy analysis. It runs all European markets simultaneously to 2050, with cross-border flows derived by replicating the EUPHEMIA algorithm used by ENTSO-E.
Outputs cover hourly prices, capture rates, negative-price depth and frequency, curtailment exposure, and BESS profitability projections at country and asset level. These are the quantitative inputs that CfD bid pricing and project bankability cases require.
Methodology
All figures are reported in hourly terms using ENTSO-E Transparency Platform data, covering day-ahead prices, day-ahead generation forecast, and actual generation series at bidding-zone level. EU day-ahead markets moved from an hourly to a 15-minute Market Time Unit on 1 October 2025. Quarter-hourly settlements from that date have been aggregated to hourly buckets using the arithmetic mean of the four 15-minute prices within each hour, so that the 2019 to 2026 series remains consistent. The EU-27 total covers EU member states represented in ENTSO-E day-ahead data, that is, EU-27 excluding Cyprus and Malta, which are not part of the Single Day-Ahead Coupling. Norway is excluded. Means and minima are simple arithmetic averages over hourly negative prices. Germany and Luxembourg share the DE-LU bidding zone and are reported jointly. Forecast-actual gaps for the April 2026 section are computed as the absolute difference between ENTSO-E day-ahead generation forecast and actual generation, by hour, for the DE-LU bidding zone.
Author: Safa Sen, Market Engagement Lead For CWE at Ricardo, Member of WSP.
Ricardo is a member of professional service firm WSP Group, uniting 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.
The views and opinions expressed in this article are the author’s own, and do not necessarily reflect those held by pv magazine.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
