PERILS, the independent Zurich-based catastrophe data provider, has published its third industry loss estimate for extratropical windstorm Nils, putting the insurance market loss at €695 million. The figure is a roughly 9% reduction from the €767 million second estimate PERILS issued three months after the event, but remains above the initial €586 million estimate published six weeks after the storm struck. The estimates cover property and motor lines of business and are based on loss data collected directly from affected insurers. A fourth and final estimate is due on February 13, 2027, twelve months after the event end date, in line with PERILS' standard reporting schedule.
Nils struck the Nouvelle-Aquitaine and Occitanie regions of southwest France between February 11 and 13, 2026, causing two fatalities and widespread disruption. The storm made landfall on the Atlantic coast on the evening of February 11 before crossing southwestern France and continuing over the Mediterranean toward Corsica, bringing destructive winds, intense rainfall and river flooding, while also generating high avalanche risk in the Alps.
Extreme winds were the primary driver of insured losses, cutting power to approximately 900,000 households and damaging thousands of houses, roofs and lightweight structures. The event followed an already unusually wet start to the year in the affected region, with ground saturation ahead of the storm contributing to subsequent flooding in the Garonne, Dordogne and Charente river basins.
Luzi Hitz, product manager at PERILS, noted that the 2025/26 European windstorm season produced two events under PERILS CORE coverage: Goretti in January 2026, now estimated at €479 million, and Nils in February, now estimated at €695 million. He added that Windstorm Kristin, which struck Portugal and Spain in late January, falls under the separate PERILS EXTENDED coverage and is currently estimated at €1,727 million - making it the largest single loss of the season by a considerable distance once that broader coverage is counted.
Hitz noted that European windstorm losses are typically dominated by events affecting the British Isles and Western Europe, making the 2025/26 season unusual in that the majority of losses fell on the Iberian Peninsula instead. For reinsurers and ILS investors whose European windstorm models are calibrated primarily on historical UK and northern European loss experience, a season concentrated in Iberia is a meaningful deviation from the expected geographic distribution.
The full movement across Nils' three estimates - €586 million at six weeks, €767 million at three months (+31%), €695 million at six months (-9%) - illustrates how substantially early post-event loss figures can move as insurers submit more complete claims data. That volatility is a normal feature of the PERILS reporting cycle rather than an indicator of unreliable data: early estimates are based on partial submissions, and the picture refines as the full loss development is reported.
For reinsurers and ILS investors using PERILS' index for parametric triggers or portfolio validation, the practical implication is that the estimate in effect at a given trigger date matters as much as the final figure. A parametric instrument whose trigger is set against the PERILS six-week estimate would have evaluated Nils against €586 million; one set against the three-month estimate would have seen €767 million. Market participants structuring instruments with European windstorm PERILS triggers should confirm at which estimate date the reference point is fixed - and whether a revision of the scale seen between Nils' first and second estimates falls within or outside the trigger band they have designed around.
For brokers and cedants placing French property risks, the CatNat context adds a further layer. Under France's natural catastrophe insurance regime, the state-backed reinsurer Caisse Centrale de Réassurance provides an unlimited guarantee for qualifying natural disaster losses. A CatNat-qualifying event changes how a French insurer's net retained loss compares to the gross PERILS industry figure - meaning the PERILS estimate for Nils should not be read as a direct proxy for any individual carrier's exposed position without accounting for the state reinsurance buffer that sits above the primary layer.
