Wildfire-exposed catastrophe bond issuance reached $5.183 billion year to date in 2026, closing on the full-year 2025 record of $5.55 billion (Artemis, August 2026).

The peril investors once refused to underwrite outright now clears the market mostly bundled inside multi-peril indemnity structures rather than sold on its own.

Key Takeaways

  • Only three of at least 20 wildfire-exposed series priced in the first half of 2026 were pure-wildfire 144A deals, so most of the $5.183 billion rides in bonds where wildfire shares a trigger with named storm, earthquake or convective storm.
  • The California FAIR Plan's debut upsized from $250 million to $750 million, the largest wildfire-exposed cat bond ever, at a 2.24% expected loss and a 9.75% spread, roughly 4.35 times expected loss.
  • Its second issuance priced at a 3.59x multiple after two upsizes and three rounds of spread tightening, from opening guidance of 9.75% to 10.75% down to a 9.5% print.
  • The compression is not only capital. Moody's RMS released North America Wildfire HD version 2.0 in October 2024, adding location-level ignition modelling and a stochastic catalog rebuilt for post-2020 severity.
  • FAIR Plan residential exposure reached $649.4 billion in June 2025, up 42% from $458.1 billion in September 2024, which is what abundant reinsurance capacity has not fixed.

The Curve, and What Is Inside It

Wildfire-exposed issuance ran $2.57 billion in 2023 and $2.84 billion in 2024 before nearly doubling to $5.55 billion in 2025, the year the January Palisades and Eaton fires generated an estimated $40 billion of insured losses. North America's aggregate natural catastrophe insured losses hit $90 billion for 2025, with secondary perils accounting for roughly 99.9% of the regional total (Swiss Re Institute, 2026).

A loss year that size might have frozen appetite at the following renewal. The year-to-date total instead sits $367 million short of matching all of 2025, with roughly five months still to run.

The composition matters more than the total. At least 20 wildfire-exposed series priced in the first half of 2026, of which three were pure-wildfire 144A deals, one behind 2025's full-year record of four. The remainder rides inside deals where wildfire shares a trigger and an aggregate expected-loss figure with named storm, earthquake, severe convective storm or winter storm.

What the Golden Bear Deals Priced

Golden Bear Re Ltd. (Series 2026-1) launched at a $250 million target and closed at $750 million, the largest wildfire-exposed catastrophe bond issued, on a 2.24% initial expected loss priced to a 9.75% risk spread, roughly 4.35 times expected loss (Artemis, 2026). The notes give the FAIR Plan indemnity, per-occurrence protection on a three-year term running to the end of 2028.

MetricGolden Bear Re 2026-1Golden Bear Re 2026-2
SponsorCalifornia FAIR Plan Association
Initial target$250 million$200 million
Final size$750 million$400 million
Initial expected loss2.24%2.65%
Priced risk spread9.75%9.5%
Multiple-at-market~4.35x3.59x
Trigger / termIndemnity, per-occurrence, 3-year

The second tranche executed harder. Series 2026-2 launched at $200 million, lifted to $350 million and then $400 million, on a 2.65% expected loss, with spread guidance opening at 9.75% to 10.75%, tightening to 9.75% to 10%, then to 9.5% to 9.75%, before pricing at 9.5% for a 3.59x multiple (Artemis, 2026). Two upsizes and three rounds of tightening on a repeat issuance in the same year is the signature of more capital chasing the peril than the sponsor needed.

Spinnaker's $100 million Mountain Re Ltd. (Series 2026-1), sponsored through Hippo's Bermuda vehicle, shows the bundled side. It covers named storm, earthquake, severe thunderstorm, winter storm and fire, with fire added for the first time and carrying the largest expected-loss contribution of any peril in the deal, ahead of named storm (Reinsurance News, 2026). It priced below guidance on an oversubscribed book, under a name that mentions no peril at all.

Market-wide softening explains part of this: cat bond pricing fell more than 20% year over year as of March 2026, with multiples roughly 30% below two years earlier.

What is wildfire-specific is calibration. Moody's RMS released version 2.0 of its North America Wildfire HD Model in October 2024, adding community- and location-level ignition modelling and a stochastic event catalog rebuilt for post-2020 severity, exposure growth in high-hazard zones and rebuilding cost inflation. "The key enabler is robust risk quantification. Investors need confidence in the underlying hazard data and catastrophe models," Marsh's Tyson Vickery said (Insurance Journal, August 2026).

Indemnity Puts the Modelling Risk on the Investor

Both Golden Bear series and Mountain Re 2026-1 are indemnity, per-occurrence structures, so recovery tracks the cedant's actual incurred losses rather than an index. That minimizes basis risk for the cedant and moves modelling risk onto the investor, who is underwriting the sponsor's claims administration and the vendor model's loss-cost calibration rather than an independently published number.

Bundling adds a second layer of the same question. A five-peril bond quotes one blended expected loss, and an investor pricing off that composite is relying on the sponsor's per-peril attribution being accurate and stable across the term. Wildfire's historical series is shorter-tailed than hurricane's and carries wider model uncertainty bands, so a blended spread that looks fair in aggregate can be thin for the wildfire slice specifically.

A cedant buying California wildfire protection inside a five-peril bond is leaning on four more mature loss-cost curves to help underwrite the least mature one. That is a reasonable trade on both sides, and it is not the same risk transfer as a dedicated wildfire tower.

The index alternative is not available yet. Verisk's Property Claim Services and PERILS AG maintain industry loss indices for European windstorm and US severe convective storm, on the premise that a neutral standardized estimate resolves disputes faster than a claims audit. No comparably granular US wildfire curve exists, because wildfire losses concentrate in narrow bands shaped by fuel load, wind corridors and suppression response, which resists the grid-based construction PERILS uses for windstorm.

The larger caution is that none of this measures primary rate adequacy. A cat bond spread is the price a diversified investor holding wildfire alongside dozens of uncorrelated exposures accepts for tail risk on a three-year note. A primary wildfire rate is what a single-state insurer, unable to diversify California's correlation, needs to cover expected loss, expense and a cost of capital.

FAIR Plan residential exposure reached $649.4 billion in June 2025, up 42% from $458.1 billion in September 2024 (Digital Insurance, July 2025), driven by non-renewals in the voluntary market, through the same period ILS capacity for the peril set records. Plenum's Dirk Schmelzer has argued wildfire is now large enough to place "on a standalone basis," a claim the three-of-twenty deal count does not yet carry, and the FAIR Plan's assessment and recoupment mechanics are what absorb the difference in a bad year.

Further Reading

Sources