By August, three things were already true this year. France had logged its biggest peacetime evacuation in modern memory. Spain had recorded the largest single wildfire in its history. And the bill for it all had blown past €3 billion before the fire season had even reached its usual August peak, according to Financial Times analysis of European Commission cost methodology. None of that should really surprise anyone who has been paying attention to EU climate science over the past decade; Brussels' own modellers flagged most of this back in 2017. What's striking is how quickly the real world is catching up with, and in places overtaking, projections that were meant to describe the end of this century.

A season that keeps breaking records

More than 250,000 hectares had burned across EU member states by late July, an area roughly the size of Luxembourg, according to satellite data from the European Forest Fire Information System (EFFIS), which runs under the EU's Copernicus programme. Fire crews had logged around 1,200 blazes by 29 July, more than double the long-term average of 569 for that point in the year. Independent fact-checking of the EFFIS data puts 2026 second only to 2022 for burned area at this stage of the season. France has set a modern national record and Spain's season already ranks among its worst.

The human cost has been severe. Fires have killed at least 17 civilians, six firefighters and two helicopter crew, and displaced roughly 330,000 people across Spain and France. The single deadliest incident, near Los Gallardos and Bédar in Andalusia, killed 14 people, including seven Britons.

The FT's analysis puts the economic cost of this year's fires and heatwaves at more than €3.1 billion for the five worst-hit eurozone countries alone, among them Portugal, Greece and Romania, in just the first two months of the season. That already outstrips the European Commission's own estimate of a €2.5 billion average annual impact across the whole bloc. ETH Zurich climate economist Sarah Meier told the FT the true toll from this year's fires could run to more than triple the roughly €2.1 billion a year she calculates Portugal, Spain, Italy and Greece lost in GDP terms to wildfire between 2011 and 2018, and that's before smoke damage to Bordeaux's wine harvest, disruption to Gironde's aerospace cluster (home to Safran, Dassault Aviation and ArianeGroup), or falling river levels forcing nuclear plant shutdowns in Hungary and Romania are fully counted.

The science saw this coming. The numbers just look worse now

Long before this summer's headlines, the European Commission's Joint Research Centre tried to put a number on exactly this kind of scenario. A 2017 technical report from the JRC's PESETA II project modelled how climate change would alter forest fire danger across Europe using the Canadian Fire Weather Index (FWI) system, the same index EFFIS still uses today to rate fire risk from daily temperature, humidity, wind and rainfall data.

What the Fire Weather Index actually measures The FWI isn't a burned-area forecast. It's a standardised score of how flammable conditions are on a given day, built from six components. Three track moisture in different fuel layers (fine litter, moderately compacted duff, and deep compacted soil), and three combine those into fire-behaviour scores for spread rate, fuel available to burn, and overall fire intensity. It only reflects weather, though. It says nothing about vegetation type, terrain or how a fire is actually fought, which is one reason actual losses can diverge sharply from what the index alone would suggest.

Feeding that index through a set of regional climate simulations, the JRC team projected that by the final decades of this century, average fire danger across five of the worst-affected southern European regions (Portugal, Spain, southern France, Italy and Greece) could climb by around 30% under a relatively high-emissions pathway. That translated into an average 97% jump in burned area across the region as a whole. Southern France came out worst in percentage terms, with a projected 184% increase, against roughly 72–93% for the Iberian peninsula and 112–121% for Greece and Italy. Under a lower-emissions pathway the increases were far smaller, closer to 20% for fire danger and well under half the burned-area growth, which shows how sensitive the outlook is to the emissions path actually taken.

Translated into money, the JRC put the current average annual restoration cost for wildfire damage across those five regions at roughly €1.97 billion a year, rising to about €4.04 billion under the higher-emissions scenario or €2.8 billion under the lower one by the century's end. Portugal was projected to remain the costliest country in both cases, with Italy overtaking Spain into second place.

The comparison isn't exact. The country groupings, time windows and cost methodologies all differ. But the direction matters: a cost trajectory drawn up to describe the 2070s and beyond is being brushed up against within a single season.

Two caveats apply to that 2017 study. First, it relied on the SRES emissions scenarios (A1B and E1) and the ENSEMBLES climate model suite, which were standard tools a decade or more ago; climate science has since largely moved on to the newer RCP and SSP scenario frameworks. The JRC's own follow-up work under its PESETA IV project, using the higher-emissions RCP8.5 pathway, reaches a similar directional conclusion: fuel moisture will decline and the zone of high fire danger will push north from the Mediterranean. So the original findings have aged reasonably well even if the specific scenario labels haven't. Second, and more relevant for insurers, the authors were explicit that their model excluded changes in vegetation and fuel load, ignition patterns and human adaptation, all factors that can push real-world losses higher or lower than a purely weather-driven model would suggest. This year's losses, arriving well ahead of the "average year" baseline, suggest those exclusions have cut the wrong way so far.

That vegetation gap is exactly what played out on the ground this summer. Daniel Bannister of the Willis Research Network has described Spain's fires as a case of "hydroclimatic whiplash": a wet, mild spring drove heavy vegetation growth, then summer drying turned it into tinder. A purely weather-based index like the FWI won't fully capture that until the fuel is already there to burn. It's also why treating the UK and southern Europe as a single correlated wildfire risk is harder than it looks. Willis's Claire Wilkinson has cautioned that grouping countries together risks oversimplifying the exposure, since risk "depends on so many factors – fuel, weather, ignition but also forest management, early detection and the availability of fire services to limit the spread."

Where this lands for UK insurers

For a London market underwriter, the interesting question isn't what happens in 2100. It's what happens at the January 2027 renewals. The signs from this summer are already shaping those pricing conversations.

Morningstar DBRS expects greater differentiation in renewal terms for loss-affected programmes or those relying on weak location data, while portfolios backed by granular geospatial data and credible mitigation should stay attractive to reinsurers. The agency reckons this year's fires will have a negative but manageable effect on the credit profiles of large, diversified European insurers. French non-life insurers reported a 95.3% combined ratio for 2025 with an aggregate Solvency II ratio of 299%, giving the sector a decent capital cushion. That's a marked contrast with last year's Los Angeles wildfires, which produced an estimated $40 billion in insured losses thanks to a dense concentration of high-value property. Spain and France have less combustible housing stock, and much of the land burning is forest or rural terrain that's often uninsured or excluded from standard cover.

That gap between economic loss and insured loss is the real story of the European wildfire market. Spain's 2025 wildfire season, its worst in three decades, with almost 355,000 hectares burned, caused close to €5 billion in economic losses, but "well under" €1 billion of that was insured, according to Marsh's global placement leader Tyson Vickery. The European Central Bank and EIOPA have separately warned that only around a quarter of losses from climate-related catastrophes across Europe between 1980 and 2024 were insured at all. Gallagher Spain's chief executive Ana Matarranz has pointed out that the Spanish market has plenty of experience handling wildfire. What's changing, she argues, is the frequency and severity of the events themselves, not the industry's familiarity with the peril.

Part of that gap comes down to modelling. Analysts have said Europe simply lacks the depth of historical wildfire data that insurers rely on to price risk accurately, a gap AXA's climate research unit has tried to help close, estimating that areas around some French cities could see nearly 70% more high-risk wildfire days a year on average by 2050.

The UK isn't insulated from any of this, and not only through European catastrophe books. Experts from Willis, speaking to Insurance Business UK after this summer's wildfires broke out on Conwy Mountain in North Wales, argue the country still lacks the modelling and risk assessment it needs as exposure grows. "I'd say the whole of the UK is under-assessed, because historically we've not really been a fire-prone country, so we don't have the data, we don't have the models really available to us to look at wildfire risk in the UK compared to places like Spain or even California," said Daniel Bannister, weather and climate risks research lead at the Willis Research Network. Unlike the US or Australia, where wildfire cover can be excluded from property policies in high-risk areas, UK insurers generally still fold it into standard fire cover. Claire Wilkinson, Willis's managing director of alternative risk transfer solutions, doubts that approach will hold indefinitely: "You can't expect insurers to keep paying claims when people just keep rebuilding in high-risk areas."

The Prudential Regulation Authority launched a General Insurance Stress Test in May 2026 to press firms on climate-driven property exposure, and domestic losses are already climbing. The Environment Agency has recorded 110 wildfires on Sites of Special Scientific Interest this year alone, part of a pattern that has seen the UK described as "underprepared" for wildfire risk even as losses hit record levels. Deloitte has separately forecast that UK home insurers will swing to a net underwriting loss in 2026, with the combined ratio reaching 102.1% as storm, flood and subsidence claims pile up. The ABI has confirmed UK subsidence claims hit £153 million in the first half of 2026, and insurers including Ecclesiastical are increasingly treating subsidence and wildfire as a combined, compounding exposure as heatwaves and dry spells become more frequent, rather than as two separate perils.

None of this is happening in a capital-starved market, at least for now. Global reinsurance capital hit a record $790 billion at the end of the first quarter of 2026, and property catastrophe buyers secured double-digit price reductions at the June and July renewals. That capacity has so far let the market absorb this summer's wildfire losses without material strain. Whether it survives contact with a wildfire that reaches a major urban area, rather than forest and scrubland, is what underwriters are now turning over ahead of 2027. Law firm Clyde & Co has observed that Europe is increasingly experiencing conditions traditionally associated with fire-prone regions like California and parts of Australia, and the pricing, data and reinsurance structures built for the old European baseline are being tested in real time.

What it means for premiums, and for brokers

For clients, broad property catastrophe pricing actually softened at the June and July 2026 treaty renewals, thanks to that record reinsurance capital. There's no market-wide wildfire shock landing on every policyholder's desk. Underneath that softening, though, DBRS's own guidance points to a widening split. Risks with strong geospatial data and demonstrable mitigation, such as defensible space, vegetation management and fire-resistant building materials, are being priced and retained on favourable terms. Loss-affected accounts, or those with thin underwriting data, are facing higher retentions and, in some cases, wildfire-specific sub-limits or exclusions creeping into European property and hospitality programmes.

For brokers, that has some practical implications. Submissions need to work harder. Clients with exposure in southern Europe, hotel groups, agricultural operators, second-home owners and tour operators, should expect underwriters to ask more granular questions about vegetation clearance, building materials and defensible space around structures than they did two or three years ago. Brokers who can bring geospatial risk data to a submission, rather than leaving underwriters to assume the worst, are likely to get better terms for their clients. Behaviour matters too. David Williams, Willis's senior director of alternative risk transfer solutions, has noted that in fire-prone markets like Australia and California, "people have been told about this risk for a very long time, and their behaviour changes during bad weather days to try to stop ignitions happening", a discipline that hasn't caught on to the same degree among UK or European clients yet.

The protection gap is worth treating as an advisory opportunity rather than just a statistic. With well under a fifth of Spain's 2025 wildfire losses insured, there's a real conversation to be had with clients who assume standard property cover extends to wildfire and land damage when it often doesn't, particularly for rural, agricultural or forestry assets. Willis's Claire Wilkinson has pointed out that tourism clients haven't yet raised wildfire specifically in cover conversations. "We've just not had the conversation yet," she said, but she thinks brokers should be proactively suggesting parametric insurance as a way to protect revenue from smoke, evacuation and dry-condition disruption, even where there's no direct property damage to trigger a traditional claim. Williams makes the economic case for that with Canada's 2024 Jasper wildfire, where insured losses reached around $1.3 billion but the fuller hit to tourism and the local economy "may not be fully realised yet". That's the kind of uninsured tail risk parametric products are designed to catch.

UK-domestic pricing pressure is a separate but related story. With Deloitte projecting a net underwriting loss for UK home insurers this year and subsidence claims already running at £153 million for the first half of 2026, brokers placing UK household and commercial property business should expect wildfire to increasingly sit alongside subsidence and flood in renewal conversations, rather than being treated as a peril that only matters overseas.

The clearest date on the calendar is the January 2027 reinsurance renewal season, where DBRS expects this differentiation to show up most visibly in treaty terms. Brokers with clients renewing around that point, especially those with meaningful southern European exposure, have a narrow window to get ahead of it with better data now.