Walk around your house with your phone for three minutes. Record the exterior, the eaves, the fence line, the vent screens, the gap between the firewood stack and the siding, and upload the video. In about the time it takes to brew coffee, an AI will return a report that identifies every combustible vulnerability on your property, ranks them by ignition risk, estimates what each fix would cost, and tells you which ones qualify for California's Safer from Wildfires insurance discounts.
That app exists, it is free, and it launched in December 2025.
Six weeks later, the Palisades and Eaton fires destroyed 13,000 homes and killed 31 people in the worst wildfire disaster in Los Angeles history, causing an estimated $31 billion in damage. What followed became the largest residential reconstruction project California has ever attempted, with governors, presidents, and local officials all competing to remove obstacles from the permitting pipeline. Among the obstacles they removed: the building codes that would have required the fixes the AI recommends.
What the AI Sees
WyldSafe, built by a Thousand Oaks engineer whose family nearly lost their home in the 2018 Woolsey Fire, uses patent-pending computer vision to analyze a homeowner's walkthrough video frame by frame. Combustible vegetation pressed against the foundation. Unscreened attic vents that funnel embers into the roof assembly. Wood fencing that acts as a wick from the property line to the exterior wall. Mulch beds within five feet of the structure. Each flagged item arrives with a remediation step and a cost estimate.
"Within 10 minutes of an ember landing on a small door, our ADU was completely engulfed," founder Nate Siggard told reporters. He built the app so homeowners could see what a fire inspector would see, privately and for free, and fix it before a citation or a wildfire forced the question.
At the insurance industry level, the analysis runs deeper. ZestyAI's Z-FIRE model processes satellite and aerial imagery, topographic data, property-level structural characteristics, and the history of more than 2,000 wildfires to produce a risk score for every property in the contiguous United States, evaluating defensible space, building materials, roof class, vent screening, and vegetation density, then cross-referencing those features against climatological and terrain exposure to predict both the likelihood of fire reaching a structure and the structure's probability of surviving it.
Z-FIRE is not experimental or aspirational. More than a third of California's insurance market uses it, including the California FAIR Plan, the state-created insurer of last resort that now covers 696,562 properties with $768 billion in total exposure. Approved for underwriting and rating in every U.S. wildfire state, it was the first AI model ever accepted as part of a California Department of Insurance rate filing.
A joint study by ZestyAI and the Insurance Institute for Business & Home Safety, examining 71,100 wildfire-exposed properties, found that homeowners who clear vegetation from the perimeter of their home nearly double their structure's likelihood of surviving a wildfire. Separate IBHS research confirmed that a five-foot noncombustible buffer around a home cuts its risk of igniting in half.
Five feet is the distance between a garbage can and a wall.
What the Code Requires
On January 13, 2025, six days after the fires started, Governor Newsom signed an executive order suspending CEQA and California Coastal Act permitting requirements for reconstruction. Subsequent orders allowed homeowners who had built to the 2019 building code to use their previously approved plans without modification, suspended building codes that would have gone into effect January 1, 2026, and waived non-safety building requirements entirely.
In June 2025, Newsom signed a bill preventing state building code updates for six years, and Los Angeles County then decided against including the Eaton Fire area in the local fire severity map, a designation that would have mandated stricter construction standards. Board of Forestry Zone Zero recommendations, which would require the kind of five-foot noncombustible perimeter that IBHS proved cuts ignition risk in half, remain in process and are unlikely to produce final requirements before the end of 2026.
None of this math is subtle. Homes that burned because they lacked defensible space, ember-resistant vents, and noncombustible siding are being rebuilt without any requirement to add those features, on the same lots, in the same fire-prone terrain, under the same wind conditions that produced the original disaster.
Urgency explains the tradeoff. Of 890 rebuild applications submitted in the Eaton Fire zone by mid-2025, only 44 permits had been issued; in Pacific Palisades, 165 out of 650. By January 2026, the one-year anniversary of the fires, fewer than a dozen homes had been fully rebuilt across Los Angeles County, with roughly 900 under construction. Permitting delays proved so severe that the state deployed AI-powered plan review tools from Steadfast LA, Autodesk, and Amazon to accelerate the process, shaving weeks off review timelines.
California is using AI to speed up permitting for homes that do not need to meet the standards AI already recommends.
Pricing the Risk Without Reducing It
Insurance markets harbor no such ambiguity about what these homes need. FAIR Plan exposure jumped 250% between September 2022 and June 2026, driven by private carriers refusing to write policies in fire-prone areas, and the plan's 29.1% rate hike starting October 15, 2026 will hit hardest in exactly the communities now rebuilding without upgraded fire-resistance standards.
California Regulation 2644.9 requires insurers to offer discounts for 12 specific home hardening measures. AB 888, the Safe Homes Act, provides grants for qualifying homeowners to implement them. AB 38 requires sellers to disclose fire vulnerabilities during home sales. Regulatory infrastructure to reward mitigation exists, but no requirement compels new construction or rebuilds in fire zones to actually implement the measures the AI identifies and the insurance framework incentivizes.
That gap is structural, not technological. An insurer can use Z-FIRE to calculate that your property has an elevated ignition probability because your vent mesh is quarter-inch instead of eighth-inch, your wood fence connects your neighbor's vegetation to your exterior wall, and your roof valley accumulates debris that ignites under ember showers, and it can price your premium accordingly, but it cannot require you to fix any of it. If you are rebuilding in the Palisades or Altadena right now, neither can the building code.
A 2025 study in Nature Communications quantified why individual compliance may not be enough even when it happens. Analyzing California's largest existing structure loss database alongside simulated fire and ember exposure conditions, researchers found that structure separation and community-wide mitigation matter as much as individual hardening, because wildfire, unlike earthquakes or floods, propagates through neighboring properties: one home with combustible landscaping can transmit fire to three code-compliant homes downwind. Parameter importance analysis showed that both defensible space and structural proximity drive survival rates, a dynamic that individual building permits cannot address and community-level codes have not yet attempted.
Speed vs. Safety
Families displaced for a year and a half, living in hotels and rental units and relatives' spare bedrooms, need to go home, and the speed-versus-safety objection deserves its full weight. Adding requirements to the permitting process, even requirements supported by strong evidence, extends timelines that are already punishing. Full wildfire hardening can cost $5,000 to $50,000 per home depending on scope, a burden that falls on homeowners who just lost everything and may be fighting their insurance company for the original claim. Newsom's calculus prioritizes speed and affordability, and there is nothing cynical about wanting people in their houses.
But that counterargument has a shelf life, because homes rebuilt to 2019 standards will stand for decades and their insurance premiums will reflect AI-assessed risk profiles that account for every hardening measure they did not install. And the next fire, burning through the same canyons with the same Santa Ana winds, will find the same vulnerabilities the AI already flagged.
What a Homeowner Can Do Right Now
If you are rebuilding in a fire zone, buying in the wildland-urban interface, or simply trying to hold onto a homeowner's insurance policy in California, the tools already exist.
Run a WyldSafe assessment, which is free, takes five minutes, and produces a prioritized list of vulnerabilities with cost estimates. Focus on the five-foot noncombustible perimeter first: move garbage cans, firewood, and planters away from exterior walls, replace combustible fencing within five feet of the structure with metal, upgrade vent screens from quarter-inch to eighth-inch mesh, and install Class A fire-rated roofing if your current roof does not already qualify. These are the measures that IBHS research shows cut ignition risk most dramatically, and they are the same measures that California Regulation 2644.9 entitles you to insurance discounts for completing.
Document everything with before-and-after photos, contractor invoices, and product specifications, because AB 888 grant applications and insurance discount claims both require proof of completed hardening, and WyldSafe generates documentation formatted for those submissions.
If you are a builder, the voluntary standard exists too. IBHS's Wildfire Prepared Home designation provides research-backed specifications for construction in fire-prone areas, covering everything from vent design to deck joist spacing to exterior wall materials, and while no jurisdiction currently requires it, every insurer will eventually reward it.
Limitations
WyldSafe's accuracy for individual properties has not been independently validated in a peer-reviewed study, and the app relies on visible exterior features without assessing interior structural components, attic insulation, or concealed vent configurations. ZestyAI's Z-FIRE model is proprietary, and while post-event analyses show strong alignment between its risk scores and actual fire outcomes, full model architecture and training data are not publicly available for independent replication. IBHS ember-resistant buffer research was conducted under controlled conditions at the institute's test facility; real-world fire behavior varies by terrain, wind speed, humidity, and fuel load in ways that no single metric can capture. Community-mitigation findings from Nature Communications apply to established neighborhoods, not necessarily to new developments where structure density and layout may differ. Finally, cost estimates for wildfire hardening vary enormously by region, contractor availability, and material costs; the $5,000 to $50,000 range cited above spans basic perimeter clearing through complete exterior material replacement, and most homeowners will fall somewhere in between.