Kettle

Kettle is a concentrated test of whether an independent wildfire model can find insurable distinctions inside areas that conventional carriers increasingly treat as a single high-hazard block. The MGA separates the problem into ignition, spread and building vulnerability, combining geospatial, satellite, weather and real-estate information to price insurance and reinsurance. Management told Insurance Journal that the system uses roughly 130 terabytes across about 40 datasets. That breadth can improve resolution, but model volume should not be confused with model validity. Wildfire is shaped by wind, fuel, drought, topography, suppression, utility infrastructure, ember transport and neighborhood-level defensibility. Those relationships change as climate, land use and mitigation evolve. A model that ranks properties well in one event can still understate absolute loss or correlation in the next. Kettle’s early products emphasized parametric cover: a pre-agreed payment can be triggered when wildfire crosses a property boundary. Parametric structures can settle quickly and avoid disputes about repair estimates. They also introduce basis risk. A fire may produce smoke, evacuation, business interruption or defensible-space expense without crossing the defined boundary; conversely, a boundary trigger may occur with limited physical damage. The boundary trigger is necessary, but it is not the only requirement. Kettle’s linked Fire in a Parcel fact sheet also requires proof of at least $10,000 of loss from the wildfire event and notice and proof within 60 days. It describes confirming the boundary intersection using National Interagency Fire Center perimeters and payment within five business days after delivery of an event report. Smoke damage without the required trigger is not covered under that description. These are company-described product conditions, not a review of an issued policy; buyers should confirm the applicable wording rather than assume that a boundary crossing alone guarantees payment. Reinsurance buyers need equal clarity about how a grid or event definition responds when one fire becomes multiple incidents or when fires merge. Kettle reported exposure to the January 2025 Los Angeles fires and later said its losses were minimal and all claims had closed by early March. It wrote about $20 million of premium in 2024, according to management. Those results provide a useful event test, not a complete validation. The fires affected specific neighborhoods and construction patterns, while a model and portfolio must survive many different wind regimes, vegetation types and accumulation shapes. Kettle itself revised its broader Los Angeles insured-loss estimate from $11 billion–$17.5 billion to $27 billion–$36 billion as the event developed. Updating with new information is appropriate, but the size of the revision illustrates epistemic uncertainty and the difference between a real-time industry estimate and a stable underwriting model. In February 2026 Kettle entered a strategic relationship with RLI to launch non-admitted multi-peril commercial property in California and Nevada, underwritten by Mt. Hawley, alongside an RLI equity investment. The move broadens Kettle from a peril-specific and parametric specialist into conventional all-risk property where wildfire is a central selection variable. It brings strong carrier paper—RLI announced an A++ AM Best rating for Mt. Hawley later that month—and experienced property underwriting. It also makes the operating model harder. Multi-peril policies require water, fire, theft, business-income, valuation and claims controls that a wildfire score cannot supply by itself. Equity ownership by the capacity provider may strengthen alignment while making independent review of model changes, rate adequacy and authority exceptions even more important. Kettle publicly describes its model and product thesis but does not disclose loss ratios, rate levels, aggregate limits, treaty structure or the economics of the RLI relationship. The measures that matter are actual-to-modeled loss by event and property cohort, calibration as well as rank ordering, total insured value and probable maximum loss by fire shed, parametric basis-risk complaints, model-version changes, mitigation credits, inspection and override outcomes, loss-estimate revisions, capacity renewal, claims speed and disputes, and performance of non-wildfire perils as the commercial book matures.

Kettle’s current commercial-property page and linked fact sheet distinguish a $200 million total insured value ceiling from a maximum $10 million policy limit. The sheet describes California and Nevada distribution exclusively through wholesale brokers, with ground-up, primary-layer and loss-limit placements. The ground-up option is described for exposures totaling up to $10 million TIV, while the loss-limit option is for risks without a mortgage. These are advertised underwriting parameters, not guaranteed capacity for any submission. MGA Index analysis: the value of an eligible property schedule is not the amount of insurance supplied. A placement assessment should identify the selected structure and any remaining uninsured or separately insured value.