About 1.5 million US homes sit in the highest ('Extreme') wildfire-risk tier, and housing in fire-exposed areas has grown 40-60%+ over the last two decades

Summary

No single, universally agreed number captures "housing units in the highest wildfire-risk areas" because the major risk-modelers use different tiers and different geographic scopes. The most directly comparable national figure is First Street Foundation's 2022 property-level wildfire model: roughly 1.5 million U.S. properties sit in its top ("Extreme," >26% chance of a wildfire event over a 30-year mortgage) risk tier, with another ~3 million at "Severe" and ~6 million at "Major" risk below that. Growth is best measured through the U.S. Forest Service's Wildland-Urban Interface (WUI) research and Cotality/CoreLogic's annual metro-level wildfire-exposure tracking: WUI housing nationally grew 41% from 30.8 million to 43.4 million houses between 1990 and 2010 (the fastest-growing land-use type in the country), and within the perimeters of fires that had already burned, houses grew 62% (177,000 to 286,000) over the same period. More recently, Cotality's metro-level tracking shows the highest-risk ("High/Extreme") housing count in California's top metros grew 35%-64% just between 2019 and 2024 (e.g., Los Angeles metro: 121,589 to 185,763 homes, +53%).

Why there is no single number

Wildfire risk to homes is modeled by several independent organizations, each with its own scoring system, risk tiers, and geographic coverage: First Street Foundation (Fire Factor, nationwide, property-level, five tiers Minor/Moderate/Major/Severe/Extreme), Cotality/CoreLogic (Wildfire Risk Score, parcel-level, western-states focus, "High"/"Extreme" combined category tracked annually by metro), and the U.S. Forest Service/academic literature (Wildland-Urban Interface, a land-use classification rather than a risk score per se, but the standard measure of housing exposure used in peer-reviewed research). None of these publish a single "housing units in the top-risk tier, tracked over 20+ years" time series, so the answer below combines the best available number from each for the pieces it is best suited to.

The current stock: ~1.5 million homes at the highest ("Extreme") tier

First Street Foundation's 2022 nationwide, peer-reviewed wildfire model (detailed in its 5th National Risk Assessment, "Fueling the Flames") found, for the contiguous United States:

• ~1.5 million properties at "Extreme" risk (greater than 26% chance of a wildfire event over a 30-year mortgage term)
• ~3 million properties at "Severe" risk (up to 26%)
• ~6 million properties at "Major" risk (up to 14%)
• ~20 million properties at "Moderate" risk (up to 6%)
• Over 49 million properties at "Minor" risk (under 1%)
• In total, over 30.4 million properties nationwide face at least a 1% chance of experiencing a wildfire over a typical mortgage term.

Cotality's narrower but more current 2025 analysis (14 western states only, "moderate-or-greater" risk, July 2025 data) counts ~2.6 million homes in that broader band across just those states, of which ~1.2 million (roughly half) carry "very high" wildfire risk — consistent in magnitude with First Street's national Severe+Extreme count of ~4.5 million, given Cotality's narrower western-only scope and different tier definitions.

The growth: housing in fire-exposed land has expanded 40%-60%+ over recent decades

National, long-run (1990-2010): the peer-reviewed USFS study by Radeloff et al. (2018, Proceedings of the National Academy of Sciences) found the U.S. wildland-urban interface (WUI) — the zone where houses meet or intermingle with wildland vegetation, and where wildfire risk to structures is concentrated — grew from 30.8 million to 43.4 million houses, a 41% increase, while WUI land area grew 33% (581,000 to 770,000 km²), making the WUI the fastest-growing land-use type in the conterminous U.S. over that period. 97% of that WUI growth came from new housing construction, not from vegetation change. Even more directly relevant to "highest risk": within the boundaries of fires that had already burned between 1990 and 2015, the number of houses present grew from 177,000 (in 1990) to 286,000 (in 2010) — a 62% increase — meaning rebuilding and new construction resumed and accelerated inside areas with demonstrated fire history.

Recent, metro-level (2019-2024): Cotality's annual wildfire-exposure analysis, which tracks homes at "High" or "Extreme" risk by metro area, shows continued rapid growth in the highest-risk stock even in the last five years:

These are the same four metro areas that ranked #1-4 for wildfire exposure in both 2019 and 2024, so the growth reflects continued building in already-recognized high-risk zones, not new areas entering the risk map. Reconstruction cost value at risk in these same metros roughly doubled over the same five years (e.g., Los Angeles: $71.0B to $143.2B).

State-level snapshots corroborating the scale

As of 2023 reporting, California had roughly 1.3 million homes classified at risk of extreme wildfires, and Colorado had more than 330,000 homes at "dire" fire risk — 12.6% of that state's entire housing stock. Cotality's 2025 state table (moderate-or-greater risk, western states only) shows California with 1,257,966 homes at that broader risk level and a combined reconstruction cost value of $796 billion, with Colorado, Texas, Oregon, and Arizona rounding out the next four states, each with over 100,000 homes at elevated risk.

Data and methodology notes

The AskAmerica/askengine connector was checked first (search_catalog for wildfire + housing-unit tables) and does carry a raw wildfire-perimeter dataset (disasters.wildfire_perimeters, NIFC WFIGS, ~39k fire polygons with acreage and cause) but has no table joining housing-unit counts to a wildfire-risk-tier classification — that is a genuine coverage gap in this connector, not a data-quality defect in an existing table, so this answer was built entirely from primary/secondary literature (First Street Foundation's own press release and methodology pages, the peer-reviewed Radeloff et al. 2018 PNAS paper via the USFS Treesearch repository, and Cotality's own published 2024 and 2025 wildfire risk reports) rather than from a connector query. Figures from different sources use different risk-tier definitions and different geographic scopes (nationwide vs. western-states-only vs. individual metro), so they are presented separately above rather than summed into one figure, and the specific tier/scope is named next to every number.