French Cal-27 baseload settled at €56.91 on 3 July, its cheapest region since 2021 and €36 under Germany. That price assumes a solid nuclear winter. The outage filings show a fleet whose availability falls into October, while the two backstops behind it — southern Norway's reservoirs and EU gas storage — are impaired at the same time. The gap between the price and the filings is the trade.
While the entire market watches Hormuz, gas storage and the injection corridor, French forward power has quietly become the cheapest winter risk in northwest Europe. Cal-27 French baseload settled at €56.91 on 3 July — a level French market commentary describes as the lowest since 2021 and explicitly attributes to “solid nuclear availability” being priced in. Germany's Cal-27 trades €36 higher. Q1-27 French baseload holds only a €12-14 premium to its own calendar.
Against that price sit four observables. Today the French nuclear fleet has 34.7 GW available out of 61.4 GW, with 6.3 GW curtailed by river temperatures on 13 July — the second curtailment episode in three weeks. The forward availability curve we derive daily from ENTSO-E outage filings shows the filed fleet falling from a late-August peak near 52 GW to roughly 45-46 GW through October: the autumn refueling campaign starts before the summer one has fully unwound. Southern Norway's reservoirs, the exporter zone behind northwest Europe's winter balance, sit 21 points below their seasonal median. And EU gas storage — the fuel for every megawatt hour a missing reactor hands to the CCGT fleet — is 51.1% full against a ~65% five-year norm.
The market is pricing the French fleet on its best recent winters, in the summer the rivers keep switching it off. Every input to the French winter premium is observable and none of them supports the comfort in the price: the filed restart queue points down into October, the slippage precedent is the historical norm, and the backstops behind a French miss are impaired simultaneously. The expression is long FR Q1-27 baseload against short FR Cal-27 — the winter premium isolated from the calendar, with no directional bet on the level of European power. If EDF lands the ramp, the premium decays a few euros. If the queue slips the way queues this shape have slipped before, the premium reprices to multiples of entry.
A calendar contract at its cheapest since 2021 is a statement about risk, not just fuel. Decomposed, the price carries four assumptions — each checkable against data that updates daily:
| Priced-In Assumption | What It Requires | Observable Status (16 Jul) |
|---|---|---|
| A solid nuclear winter | ~50-55 GW available through Dec-Feb, restarts on schedule | Filed October fleet 45.4 GW, below September; ~10+ GW of returns must land in the 7 weeks after the filing window |
| Summer curtailments stay a summer story | Heat outages don't disturb the maintenance schedule | 6.3 GW curtailed 13 Jul (St-Alban 1-2, Bugey 3-5, Golfech 2, Blayais 1+3); second episode in 3 weeks, interleaving with planned work |
| Imports backstop any miss | Nordic surplus + comfortable neighbours | NO2 reservoirs 50.7% vs 72% median; Montel floats a Nordic hydro balance near −27 TWh by end-September, weakest in ~20 years; Oslo drafting a reservoirs-before-exports mechanism |
| Gas backstops the backstop | Cheap CCGT fuel for scarcity hours | EU storage 51.1% vs ~65% norm; injections lag the 80% corridor; every replacement cargo bid away from Asia |
None of these four checks is exotic. What is unusual is that nobody is running them together. The gas market is consumed by Hormuz and the injection corridor. The power market has internalised two calm French winters and a record export year, and reads July's curtailment headlines as a heat story with an EDF reassurance attached. The reassurance — no threat to supply security — is about July. The price extends it to January.
The chart below is not a forecast. It is the French nuclear fleet's filed availability: fleet capacity minus every planned and forced outage EDF has filed to ENTSO-E, aggregated daily by Voltstack from the live outage feed across a rolling 120-day window. It re-derives every day as filings are added, extended or withdrawn — which is exactly what makes it a tradeable observable rather than a narrative.
Three features matter. First, today's trough: 34.7 GW available, the heat curtailments stacked on top of scheduled work. Second, the August plateau near 50-52 GW — the fleet's best filed window of the next four months lands in the season with the least demand for it. Third, and this is the thesis: the curve turns down into October. The filed October fleet averages 45.4 GW, below the September average, as the autumn refueling campaign opens. The fleet does not build toward winter. It walks into winter's doorway losing altitude.
To deliver the ~50-55 GW winters the price assumes, EDF must land roughly 10 GW of net returns in the seven weeks between the end of the filing window (mid-October) and the December demand ramp — the steepest scheduled restart ramp since 2022, executed in the exact weeks where French restart schedules have historically slipped. And the heat is not a side story: every curtailment day at Bugey or Golfech is lost margin for pre-winter maintenance, compressing the same autumn calendar from the other side.
The curve shows filed outages inside a 120-day window. Unfiled extensions would make October worse; newly filed accelerations would make it better. Both show up in the same feed within a day of filing, which is why the invalidation criteria below are written against the filings themselves. October averaging below September is partly just the shape of a normal autumn campaign — the point is the level it starts from, the size of the ramp it implies, and the zero slack left if any of it slips.
Autumn 2022 is the reference case for what a congested French restart queue does to winter contracts. The mechanism was not the stress-corrosion defect itself — it was the serial repricing of a schedule everyone could see:
| 2022 Marker | Promise | Reality |
|---|---|---|
| Fleet availability at the low | — | ~40% of fleet |
| 2022 output guidance | 295-315 TWh | cut to 280-300, then 275-285 TWh |
| Winter restart plan | full fleet by winter | 27 reactors by end-Dec; 5 slipped to Jan-Feb |
| FR Q1-23 baseload | ~€200 in spring | peaked above €1,000/MWh (Aug 2022) |
Nobody needs 2022's crisis to repeat for this trade to work — scenario A below assumes nothing but ordinary slippage, the kind French autumn campaigns have produced in most years since 2016. The 2022 episode defines the tail, not the base case. What it teaches about the base case is the direction of surprise: when a restart queue is congested and the buffer around it is thin, schedule news breaks one way.
A French nuclear miss is survivable when the systems around it have slack. This year the slack is the anomaly:
This is the winter face of the argument we published in June as The Correlated Catastrophe: hydro, nuclear and thermal capacity share weather drivers and fail together. July has already validated the summer half — the same heatwave curtailed the French fleet, drained nothing back into Norwegian reservoirs, and produced the redispatch spike. The winter half is what FR Q1-27 is refusing to price.
Probabilities and levels are Voltstack's analytical judgement, anchored to FR Cal-27 near €57 and FR Q1-27 near €70 (approximate mid-July levels — verify at entry; the premium, Q1-27 minus Cal-27, is the traded quantity).
Two to four units slip two to six weeks each through the October-December ramp — the pattern of most French autumn campaigns, no crisis required. RTE's winter outlook (November) flags vigilance windows; the market re-prices a French winter it had marked as solved while storage lands below target and NO2 enters winter below median.
| Marker | Now | Target (by Dec-26) |
|---|---|---|
| FR Q1-27 baseload | ~€70/MWh | €85–100/MWh |
| Winter premium (Q1-27 − Cal-27) | ~€13 | €25–35 |
Implication: the premium doubles on schedule news alone. Normal weather, normal slippage.
Slippage meets a cold spell. French demand rises roughly 2.4 GW per degree of cold (RTE); the import cushion is thin because NO2 is defending reservoirs and Germany is burning stored gas it does not have. Spot scarcity feeds the front of the curve; Q1-27 converges up toward it. The 2022 tail (Q1-23 above €1,000) frames how far this branch can run — the target below deliberately assumes only a fraction of it.
| Marker | Now | Target |
|---|---|---|
| FR Q1-27 baseload | ~€70/MWh | €120–180+/MWh |
| Winter premium (Q1-27 − Cal-27) | ~€13 | €50–100+ |
Implication: the premium pays a multiple of entry. This is the tail the €57 calendar is giving away.
The filings hold: October's returns land on schedule, no unit slips more than a fortnight, heat curtailments end with August, and a mild early winter lets the premium bleed back toward carry. The comfort in the price turns out to be earned.
| Marker | Now | Target |
|---|---|---|
| FR Q1-27 baseload | ~€70/MWh | €60–65/MWh |
| Winter premium (Q1-27 − Cal-27) | ~€13 | €5–8 |
Implication: the premium compresses toward its structural floor, not zero — French thermosensitivity means Q1 does not trade through the calendar. Risk ~€5-8 against €12-20+ in A and €35+ in B.
Buy FR Q1-27 baseload, sell FR Cal-27 baseload in equal MW. The position is long the French winter premium and nothing else: the level of European power, the gas price and the carbon price largely cancel across the legs. The short leg is the instrument carrying the comfort — Cal-27 at its cheapest since 2021 is precisely where the “solid nuclear winter” assumption lives. Entry reference: premium near €12-14/MWh (verify at entry). The premium's structural floor under French winter thermosensitivity bounds the loss; the 2022 tail bounds nothing on the other side.
Relative-value variant: long FR Q1-27 vs short DE Q1-27 for desks that want the France-specific leg against the tightest neighbour — the spread that inverted violently in 2022 whenever French scarcity set the marginal price.
It is not a cold-winter bet: scenario A pays on schedule slippage with normal weather. It is not the 2022 trade re-run: no new defect is assumed — the queue in the filings is enough. And it is not an options structure this time, stated plainly: EEX French power options do not offer the liquidity to own this in convexity at reasonable cost, so the isolation spread carries the view and the invalidation discipline replaces the defined-risk premium. Desks with access to OTC FR optionality can substitute Q1-27 call structures and keep everything else in this note unchanged.
| Signal | What It Tells You | Read |
|---|---|---|
| The D+120 filed availability curve | The queue itself, re-derived daily from ENTSO-E filings. The chart above is this feed. | The single most important input |
| Filing revisions (the slip counter) | Outage documents are versioned; every extended return date is a filed, timestamped slip. | Slips before October = early confirmation |
| River temperatures / curtailment notices | Whether the heat keeps eating pre-winter maintenance margin through August. | Extensions past mid-August compress the ramp |
| NVE weekly NO2 fill vs median | The import cushion. Live on the Voltstack Nordic hydro feed. | Flat or falling deviation = confirming |
| AGSI+ injection corridor | The price of the gas backstop. Live on the storage tracker. | Pace below 0.27 pp/day = confirming |
| FR net export flips (ENTSO-E flows) | France importing on summer evenings is the early tell that the balance is tighter than the calendar price implies. | Context, not trigger |
The filed October window holding below ~47 GW on each weekly re-derivation, or any autumn return date extended in a filing revision. Slips filed before October are the earliest tradeable confirmation.
River-temperature curtailments extending past mid-August, or new forced-outage filings at units scheduled to anchor the December fleet.
NO2 fill deviation vs median failing to close before freeze-up, Norwegian export-restriction policy advancing, or the AGSI+ corridor still lagging the 80% path into September.
RTE winter outlook (November) naming vigilance windows for December-February, or EDF trimming output guidance the way serial guidance cuts began in 2022.
Filings pull returns forward so the filed October fleet recovers above ~50 GW, and September actual availability tracks the filed curve within ~1 GW. The queue is clearing; the comfort was earned.
By end-September: no filed slips, NO2 recovering toward its median, and injections back on the 80% corridor. All three backstops healing removes the asymmetry; exit before the premium finishes decaying.
Size to the premium compressing to its structural floor in Scenario C, not to the Scenario B payoff. Take profit if the premium reprices to the Scenario A target without filed slippage to justify holding for more. Reassess on every weekly curve re-derivation and every filing revision. The invalidation is document-based and observable: this thesis is falsified by filings, not forecasts.
“Cal-27 France at its cheapest since 2021 says the nuclear problem is solved. The filings say October has less nuclear than September.” One of those two statements re-derives from official documents every morning. We are long the one that does.
This thesis exists because of a data gap: everyone can see today's French availability, and almost nobody maintains the forward view the filings contain. Closing gaps like that is what Voltstack does.
| Capability | Role in This Thesis | Generic Alternative |
|---|---|---|
| FR nuclear D+120 availability curve | The queue, derived daily from ENTSO-E outage filings (95 nuclear units, versioned revisions). The chart on this page is the live feed. | Reading REMIT filings by hand |
| NVE Nordic hydro bands | NO2 fill vs 30-year median/min bands, weekly — the import cushion, quantified. | NVE spreadsheets, no percentile context |
| AGSI+ storage corridor | The gas backstop's price, live against the 1 November target. | Weekly CSV from GIE |
| Redispatch & grid-stress feed | The correlated-stress meter: the 2 July tenfold print came out of this feed, measured, not estimated. | netztransparenz portal, per-TSO CSVs |
| Curve & spread monitors + REMIT II audit trail | The Q1/Cal premium tracked with alerts at entry and invalidation; the rationale for entering, holding and cutting exists as a timestamped by-product. | Excel and screenshots |
Outage-filing availability curves · NVE hydro bands · AGSI+/ALSI+ live · JAO flow-based domain · Redispatch & balancing stress · REMIT II native
voltstack.energy