FM Acquires FortressFire to Enhance Wildfire Risk Management Solutions

FM’s acquisition of FortressFire is more than a technology deal. It is a clear signal that the next phase of property insurance will increasingly depend on understanding exactly why one building survives a wildfire while another does not.

Commercial property insurer FM announced the acquisition of FortressFire on August 10, 2026, bringing a structure-specific wildfire intelligence company inside one of the insurance industry’s best-known engineering organizations. FortressFire will operate as an independent, wholly owned division of FM while retaining its brand and leadership. Financial terms were not disclosed.

For agents, brokers and carriers, the important part is not simply that another insurer bought an analytics company. It is what the combination says about where wildfire underwriting is heading. Broad geographic risk scores are increasingly being supplemented by much more detailed information about individual structures, surrounding vegetation, construction characteristics, ignition pathways and mitigation measures.

That could gradually change conversations about underwriting, pricing and insurability. Instead of asking only whether a property sits in a wildfire-prone ZIP code, the industry is moving toward a more practical question: What has actually been done to make this specific property less likely to burn?

Why FortressFire Fits FM’s Loss Prevention Model

The acquisition is particularly notable because of FM’s long-standing approach to commercial property insurance. Nearly two centuries old, the mutual insurer has built much of its identity around the idea that property losses can often be prevented through engineering rather than simply financed after they occur.

FM says its risk management organization includes more than 2,000 engineers working across 14 countries. Its engineers and researchers analyze property hazards, visit facilities, test loss scenarios and recommend practical improvements designed to reduce both the frequency and severity of losses. FM also serves many of the world’s largest organizations, including approximately one in four Fortune 500 companies.

FortressFire adds another layer to that approach. Its wildfire intelligence platform combines machine learning with physics-based modeling to evaluate how wildfire could interact with an individual structure. Its services include aerial property reports, analytics, inspections, mitigation assessments and ongoing monitoring. The company describes its AMP platform around four concepts: Assess, Mitigate, Monitor and Protect.

“FortressFire shares FM's core belief in the power of data-driven, location-based risk mitigation and protection measures.”

That philosophical alignment matters. FortressFire is not simply trying to predict where wildfires might occur. Its approach focuses on whether a particular structure is likely to ignite and what could be changed to reduce that vulnerability.

Wildfire Losses Are Making Property-Level Detail More Valuable

The economics surrounding wildfire make that distinction increasingly important. Swiss Re Institute reported that global insured natural catastrophe losses reached $107 billion in 2025. Secondary perils, including wildfire, severe convective storms and floods, accounted for a record 92 percent of those losses. The Los Angeles wildfires alone generated an estimated $40 billion in combined insured losses, making them the largest insured wildfire loss event in Swiss Re’s records.

The trend is not confined to one disastrous fire season. Swiss Re Institute estimates that global insured wildfire losses have been growing at roughly 12 percent annually in real terms since 1970. In North America, the growth rate is more than 14 percent annually, and Swiss Re estimates that a substantial portion of that increase cannot be explained simply by having more buildings and insured value in exposed areas. Changes in hazard and vulnerability also matter.

Even a relatively quiet catastrophe period can therefore be misleading. Swiss Re estimated insured natural catastrophe losses of $42 billion during the first half of 2026, below the 10-year first-half average. At the same time, it warned that wildfire remains the fastest-growing weather peril globally and that longer fire seasons, heat, dry conditions and continued development in exposed areas are maintaining pressure on the risk.

“A less costly first half of the year does not mean the risk has gone away.”

The Bigger Shift: From Wildfire Zones to Individual Buildings

Traditional catastrophe modeling remains essential for understanding portfolio exposure. But the industry is increasingly interested in what happens after a broad model identifies a risky area.

Two buildings on the same road can have dramatically different vulnerabilities. Roof materials, vents, decks, fences, landscaping, surrounding fuel, nearby structures and maintenance practices can all affect how fire and embers interact with a property.

Research from the Insurance Institute for Business & Home Safety reinforces that point. IBHS has emphasized the importance of building hardening and maintaining a noncombustible area within the first five feet surrounding a structure. In 2025, the organization reported research showing that an ember-resistant buffer around a home can cut ignition risk substantially. Its broader wildfire guidance also addresses roofs, vents, decks, fencing, vegetation and accessory structures as interconnected components of wildfire resilience.

Although much of that research focuses on homes, the underlying concept carries directly into commercial property discussions. Wildfire risk is not merely a characteristic of a location. It is a combination of hazard, exposure and vulnerability, and vulnerability is the component property owners can often influence most directly.

Why Mitigation Data Could Become More Important in Underwriting

For years, one of the difficult questions surrounding wildfire mitigation has been whether insurers can consistently measure it and incorporate it into underwriting. A property owner may say vegetation has been cleared, vents have been upgraded or combustible materials have been removed, but an underwriter needs reliable information to determine whether those improvements materially changed the risk.

Structure-specific analytics can help close that gap. If mitigation can be identified, documented, monitored and connected to an expected reduction in ignition probability, it becomes more useful to underwriting teams and potentially more meaningful in discussions about eligibility, pricing, deductibles and capacity.

That does not mean every mitigation investment will automatically produce a premium credit. Insurance pricing still depends on carrier appetite, regulatory requirements, portfolio concentration, reinsurance costs and many other considerations. But better property-level evidence gives insurers more information with which to distinguish a mitigated risk from an otherwise similar unmitigated property.

Regulators Are Also Moving Toward More Sophisticated Models

The shift is occurring alongside changes in insurance regulation. California, the market at the center of many wildfire availability debates, has moved toward allowing forward-looking catastrophe models to play a greater role in property insurance ratemaking while tying those reforms to commitments intended to increase insurance availability in wildfire-distressed areas. The state has also established review requirements for wildfire catastrophe models used in the regulatory process.

More broadly, artificial intelligence and predictive modeling are receiving increasing regulatory attention throughout the country. APCIA’s tracking of state insurance regulation shows that numerous states have adopted or adapted the NAIC’s framework governing insurers’ use of artificial intelligence systems. That scrutiny means carriers will need not only sophisticated models, but also strong governance, validation and documentation around how those models influence insurance decisions.

In other words, the opportunity is not simply to collect more data. The competitive advantage will come from turning reliable data into explainable, defensible and actionable risk decisions.

What Agents and Brokers Should Be Watching

For insurance professionals working with properties in wildfire-exposed areas, the FM transaction is another reason to move mitigation conversations earlier in the renewal process. The strongest submission may increasingly be one that explains not only the exposure, but also what the insured has done about it.

• Document mitigation: Maintain records of vegetation work, inspections, roof improvements, vent upgrades and other property improvements.
• Start early: Address wildfire recommendations well before renewal rather than after an underwriter raises concerns.
• Ask better questions: Find out whether carriers recognize property-level mitigation when evaluating eligibility, terms or deductibles.
• Connect teams: Bring risk managers, facilities personnel, engineers and insurance decision-makers into the same conversation.
• Monitor continuously: Treat wildfire resilience as ongoing property maintenance rather than a one-time project.

This is especially important for commercial accounts with multiple locations. A portfolio-level wildfire score can identify concentration, but location-level information can help determine which facilities deserve immediate investment and which improvements may have the greatest loss-prevention value.

Wildfire Analytics Is Becoming a Competitive Battleground

FM is not alone in seeing strategic value in more sophisticated wildfire intelligence. In late 2025, FutureProof Technologies announced an agreement to acquire Terrafuse AI, combining Terrafuse’s wildfire prediction technology with an underwriting and pricing platform focused on catastrophe-exposed property.

The transactions point toward broader consolidation between insurance organizations and specialized climate analytics providers. Catastrophe data is moving closer to the underwriting transaction itself, where insights can potentially affect risk selection, pricing, capacity and mitigation recommendations in real time.

That development should interest traditional carriers as much as insurtechs. Models alone are becoming easier to access. What is harder to replicate is a combination of high-quality data, scientific expertise, engineering knowledge, field inspection capabilities, underwriting experience and a large enough portfolio to continually learn from real-world outcomes.

The Opportunity Is to Turn Better Data Into Better Risk

For the industry, perhaps the most significant element of the FortressFire acquisition is its focus on prevention. Property insurance has traditionally excelled at transferring financial risk. Increasing catastrophe pressure is pushing carriers, brokers and insureds to spend more time on the physical risk itself.

That is particularly relevant in wildfire markets, where increasingly precise models could otherwise be used simply to identify properties insurers do not want. The more constructive opportunity is to use those same tools to identify why a property is vulnerable and what can realistically be improved.

If insurers can reliably connect mitigation actions with measurable reductions in expected loss, the conversation changes. A wildfire model becomes more than a screening mechanism. It becomes a roadmap for improving the risk.

For agents and brokers, that creates a valuable advisory role. Clients will still need help finding capacity and negotiating terms, but they may increasingly need guidance on the property improvements, documentation and ongoing maintenance that make their risks more attractive to the insurance market in the first place.

FM’s acquisition of FortressFire reinforces a direction the property insurance industry has been moving toward for years: more granular data, more engineering, more continuous risk monitoring and a stronger connection between what a property owner does before a wildfire and what insurance options may be available afterward.

For an industry wrestling with wildfire affordability and availability, the most promising technology may not simply be the model that predicts the next loss. It may be the one that helps prevent it.