The Wildfire Risk Modeling Landscape
Wildfire catastrophe modeling has become one of the fastest-evolving segments in insurance analytics. The 2017-2018 California wildfire seasons, the 2020 record-breaking fire year, and the devastating 2025 Los Angeles firestorm have fundamentally changed how insurers approach wildfire risk—shifting the industry from backward-looking actuarial tables to forward-looking probabilistic and AI-driven models.
Why Traditional Approaches Failed
Before modern wildfire CAT models, insurers relied on historical loss averages and simple wildfire hazard scores. This approach systematically underpriced risk in areas experiencing rapid vegetation change, drought intensification, and WUI (Wildland-Urban Interface) expansion. The result: insurer insolvencies, market exits from fire-prone states, and a growing protection gap.
Major Model Categories
- Probabilistic Catastrophe Models
- Companies like Moody's RMS, Verisk, and Karen Clark & Company offer full stochastic event sets with hundreds of thousands of simulated wildfire scenarios. These models estimate expected losses, tail risk, and return periods for portfolio-level decisions. Moody's RMS North America Wildfire HD v2.0, released in October 2024, explicitly simulates urban conflagration—the house-to-house fire spread that drove catastrophic losses in the LA fires.
- AI-Driven Property-Level Scoring
- ZestyAI's Z-FIRE and Cape Analytics use aerial imagery, machine learning, and property-specific features (roof material, defensible space, vegetation proximity) to produce granular risk scores. Z-FIRE is approved for rating in all major wildfire-prone states and is used by over one-third of California's insurance market.
- Physics-Based Fire Behavior Models
- Technosylva (acquired by Lockheed Martin in 2024), Pyrologix (now part of Vibrant Planet), and KatRisk simulate actual fire ignition, spread, and suppression dynamics using fuel maps, topography, and weather conditions.
Regulatory Tailwinds
California's 2025 regulatory changes eliminated longstanding restrictions on using catastrophe models for ratemaking. Insurers can now incorporate forward-looking models and reinsurance costs into pricing—a fundamental shift that has accelerated demand for approved wildfire CAT models. Moody's, Verisk, and KCC have all completed or submitted models for California Department of Insurance review.
Climate Change Integration
Leading models now incorporate climate projections. KCC's Wildfire Reference Model includes 9 climate change scenarios across 3 time horizons (2025, 2030, 2050). Kettle's AI platform ingests over 130 terabytes of satellite, weather, and real estate data to simulate millions of wildfire scenarios under changing climate conditions.