The 2026 National Roof Age Report: How Old Are America's Roofs?

What does the 2026 National Roof Age Report find?

Roofs make the news when a hurricane peels them off. The larger story is quieter and slower: the United States is living under an aging roof stock because it is living in an aging housing stock, and the two biggest building booms in modern American history (the early 1970s and the years around 2005) are both sending waves of roofs toward replacement age in this decade. That demographic fact, more than any single storm season, is why replacement demand has stayed strong, why insurers keep tightening roof terms, and why "how old is your roof?" has become a question homeowners get asked by companies they have never called.

This report proceeds like a report should. First, the housing-stock numbers and where they come from. Second, the logic that converts house age into roof age, including the Cohort Clock, a simple framework for turning a build year into a realistic first guess about a roof. Third, the geography: which regions and states have the oldest homes, why roof age and house age pull apart, and a chapter-by-chapter tour of six regions. Fourth, the two forces compounding the aging story: storm losses and insurance underwriting. Fifth, what our own roof imagery corpus shows about roofs turning over, and what a corpus like it cannot show. Finally, what a homeowner should actually do with any of this, an honest accounting of what this report does not claim, and a set of citation-ready lines for anyone quoting it.

How old is America's housing stock in 2026?

Start with the houses, because the roofs sit on them. The National Association of Home Builders, analyzing the Census Bureau's American Community Survey, puts the median age of an owner-occupied home in the United States at 42 years as of 2024, up from 31 years in 2005. Read that sentence again, because the change is the story: in less than two decades, the typical American owner-occupied home aged eleven years. Half of the country's more than 85 million owner-occupied homes are now older than 42; a home built in 1984, the year the first Macintosh shipped, is the national midpoint.

The distribution behind that median comes from the ACS's year-structure-built tabulations (table B25034). In round numbers: roughly 47 percent of owner-occupied homes were built before 1980, about a third before 1970, and only about an eighth since 2010. The shares drift a point or two between survey years, but the shape has been stable for a decade: a big old base, a large middle from the 1980s-to-2000s building runs, and a thin new layer on top.

Why is the stock aging so fast? Because the country stopped renewing it. After housing starts collapsed in 2008–2009, the 2010s produced the fewest new homes of any decade in the modern Census construction record, and starts in the 2020s (running roughly 1.3 to 1.6 million a year) have not closed the gap. When few new homes enter the pool, the median age of the pool climbs almost mechanically, the same way a town's median age climbs when no children are born. Harvard's Joint Center for Housing Studies has made the same point from the remodeling side: an older stock needs more repair and replacement spending simply to stand still, and roofing sits near the top of that spending.

Why does an aging housing stock mean an aging roof stock?

A house lasts a century or more; its roof does not. Most American homes are covered in asphalt shingles (industry estimates generally put asphalt on roughly three-quarters or more of U.S. homes), and an asphalt roof typically gives 15 to 30 years of service depending on grade, climate, and installation quality. Metal, tile, and slate last longer, but they cover a minority of houses. So the arithmetic of an aging housing stock is straightforward: a house built in 1975 is not sitting under a 50-year-old roof. It is almost certainly on its second roof and quite possibly its third, and each of those replacements happened on a schedule set by the original build date plus one or two roof lifetimes.

That is why the useful unit of analysis for a national roof report is not the individual house but the building cohort. When a large group of homes goes up in the same few years, their original roofs come due together, and so does each replacement wave after that, blurred a little by material choices and storm strikes but recognizably a wave. The 42-year median home age means the center of the American housing stock is now two full asphalt lifetimes from its build date. A country in that position replaces a lot of roofs every year whether or not the weather cooperates.

The center of the American housing stock is now two full asphalt lifetimes from its build date.

The cohort effect is visible across image captures. In the historical imagery behind this report, the most consistent pattern is that homes built together get re-roofed in clusters: a tract subdivision's roofs change color and texture within a few years of one another, first around the two-decade mark and again a couple of decades later. We present that as a pattern rather than a number, but it is exactly what the cohort arithmetic predicts, and it is why a home's build era is a legitimate starting clue about its roof even when no paperwork survives.

Two cohorts matter most right now. The first is the early-1970s boom, whose homes are on their second or third roofs, with many due for another replacement decision this decade. The second is the late-1990s-to-2006 boom (the largest building run since the first), whose original roofs are now roughly 20 to 28 years old, squarely inside the window in which asphalt roofs are commonly replaced. The next two chapters give you the tools to place any specific house against those waves: first the Cohort Clock, then the construction history that drives it.

The Cohort Clock: turning a build year into a roof-age guess

The Cohort Clock is the framework this report uses to turn a build year into a realistic first guess about a roof. It rests on three inputs: when the house was built, how long the dominant roofing material in its market typically lasts, and how many full replacement cycles fit between the build year and today. For asphalt shingles, use a cycle of roughly 20 to 25 years — long enough to reflect real service lives, short enough to reflect the fact that many roofs are replaced before outright failure. Count forward from the build year in those cycles, and wherever 2026 lands inside the current cycle is your first guess about the roof: early, middle, or late in its life.

1. Find the build year. It is on your county assessor's property record, your closing documents, and most listing sites. If you know only a decade, use its midpoint.
2. Count forward in 20-to-25-year cycles until you pass 2026. A 1975 house passes cycle marks around 1995–2000 and 2015–2025, so it is likely on its third roof, or late in its second.
3. Place 2026 inside the current cycle. A 2002 house has passed no full cycle: its original roof is 24 years old and at or past the typical replacement point. A 2016 house is 10 years in: early-to-mid cycle.
4. Then correct the guess with real evidence; our roof lifespan guide walks the records hunt, from invoice and permit down to imagery history.

Two honest limits. First, the Cohort Clock assumes asphalt; a tile, metal, or slate roof runs a much longer material cycle, though tile underlayment quietly runs an asphalt-like clock of its own. Second, the Clock assumes wear retires the roof. But across the hail belt and the hurricane coast, storms retire roofs early and reset the cycle at a date no arithmetic can predict, which is why the Clock is a starting guess, never a conclusion. For what actually determines where in the 15-to-30-year asphalt range a specific roof lands, see our guide to how long roofs last.

Building booms become re-roofing waves

If roofs are replaced in cohorts, then the history of American homebuilding is also a forecast of American re-roofing. The Census Bureau's New Residential Construction series, which tracks housing starts back to 1959, shows a landscape of booms and busts rather than a steady line. Starts ran high through the postwar decades, peaked at about 2.36 million in 1972 (still the record annual total), fell hard in the early-1980s interest-rate shock, recovered through the 1990s, climbed past two million again in 2005, and then collapsed to roughly 550,000 in 2009, the lowest annual total in the series. The 2010s recovery was slow; the 2020s have run between roughly 1.3 and 1.6 million starts a year.

Each of those swings created a cohort, and each cohort throws off replacement waves at roughly 20-to-25-year intervals: an original-roof wave, then an echo when the replacements themselves wear out, then a fainter echo after that. The table below runs the arithmetic. It is illustrative, not measured. Storms pull replacements earlier, long-lived materials push them later, and every wave spreads out as it repeats. But the shape it sketches matches what the roofing trade has experienced for decades: busy stretches that track the booms of one roof-lifetime ago.

Notice what the arithmetic implies about the next several years. The largest cohort since the early 1970s is inside its original replacement window right now, and the 1970s cohort's third-roof echo is arriving on top of it. Meanwhile the thin bust-era cohort guarantees a lull afterward. If you own a home built between the late 1990s and 2006, you are part of the single biggest re-roofing wave of the decade, competing for the same crews, materials, and insurance goodwill as millions of neighbors on the same clock. Getting ahead of a wave is cheaper than riding its crest; the budgeting chapter below returns to this.

One more wrinkle worth naming: replacement waves and building booms happen in different places. The 2000s boom was concentrated in the Sun Belt (Texas, Florida, Arizona, Nevada, Georgia, the Carolinas), so the current original-roof wave is disproportionately a Sun Belt event, landing in exactly the markets where storm exposure and insurance tightening are already most intense. The regional chapters take that up in detail.

Which regions have the oldest roofs?

The oldest roofs, like the oldest houses, are concentrated in the Northeast, and the youngest roofs are in the storm-swept center and south of the country. In the milder-weather parts of New York, New England, and the upper Midwest, roofs are mostly retired by wear rather than by storms, so roof ages there tend to track the region's old build cohorts, and asphalt roofs routinely serve into their twenties. In the hail-prone Southern Plains and Front Range and along the hurricane-exposed Gulf and Southeast coasts, storms retire roofs a decade or more early, so those regions carry the country's youngest roof stock even where the houses are middle-aged. That is the answer; now the caveat.

No one has directly counted roof ages, and what public data measures is house age. NAHB's state-level analysis of Census data puts the median owner-occupied home in New York above 60 years old, with Massachusetts and Rhode Island close behind, while Nevada's median home is in its mid-twenties and Texas's in its late twenties. The regional roof-age reading above is an inference built on that house-age data plus the replacement logic of the previous chapters, and it can be wrong for any particular street. That is why this report states it as a tendency, checks it against imagery, and never attaches invented percentages to it.

The storm effect shows up from overhead, too. In hail-prone metros, streets commonly display a patchwork of recently replaced and original roofs, while in milder metros the roof ages on a given street track the build cohort much more closely. We note that as an observed pattern rather than a statistic; imagery coverage and capture frequency vary by metro, and a partial re-roof can look like a full one across image captures. The six regional chapters below walk through each region's picture in detail, and our regional roof lifespan atlas covers what each climate does to a roof mechanically.

The Three Divergence Forces: why roof age pulls away from house age

If roof age simply followed house age, this report could end with the Census tables. It cannot, because three forces pull the two apart, and the size of the gap between a region's house ages and its roof ages is itself informative. This report calls them the Three Divergence Forces, and the regional chapters that follow use them as a common lens: for each region, ask which force dominates, and the local roof-age picture usually falls into place.

Storm turnover is the strongest force and the easiest to misread. A region where roofs average young is not necessarily a region where roofs are in good shape; it can be one where roofs keep getting destroyed. The Insurance Information Institute's claims data makes the point from the carrier side: wind and hail are consistently the most frequent cause of homeowners insurance claims, which means a substantial share of American re-roofing is scheduled by the atmosphere. In those markets, the Cohort Clock still tells you which homes carry original boom-era roofs, but a single spring storm can advance thousands of clocks to zero in an afternoon.

Material mix works in the opposite direction. The Northeast's oldest housing includes slate and standing-seam metal that has outlived several generations of asphalt; the Southwest and Florida carry large tile inventories whose surfaces last half a century even though the underlayment beneath them does not. Where these materials are common, house age overstates roof risk. And climate pace means even a correctly measured roof age translates differently by region: a 15-year-old asphalt roof is late-middle-aged in Phoenix, ordinary in Philadelphia, and, if the moss has been kept off, comfortable in Portland. Keep all three forces in hand as you read the regional chapters; each one is a different mix of the same physics.

Which states have the oldest and youngest homes?

State-level house-age data sharpens the regional picture, and NAHB's analyses of Census American Community Survey data tell a consistent story year after year. The oldest housing stock in the country is not in a state at all: the District of Columbia's median owner-occupied home predates most of the modern suburbs. Among the 50 states, New York is the oldest, with a median owner-occupied home age above 60 years, and Massachusetts and Rhode Island follow close behind. Southern New England and the Mid-Atlantic built early, built densely, and have added comparatively little since. At the other end, Nevada's median owner-occupied home is in its mid-twenties, with Texas in its late twenties and states like Arizona, Utah, Idaho, Georgia, and the Carolinas clustered among the youngest, the footprint of the very booms the wave chapter described.

Two readings of this table matter for roofs. First, the old states are old because they stopped adding homes, not because their homes are failing: a 60-year-old New York house has typically been re-roofed twice and may well carry material chosen for the climate a century ago. Second, the young states are precisely where the 2000s cohort concentrates, which means the states with the youngest houses are, paradoxically, facing some of the heaviest near-term replacement demand, because their enormous original-roof cohort is hitting the 20-to-25-year mark together. Far from insulating Nevada and Texas from the aging story, their youth puts them at the leading edge of its next chapter.

What does roof aging look like in the Northeast?

The Northeast is the old-house capital of America (New York the oldest of the 50 states by median home age, southern New England close behind, and city neighborhoods where the housing predates the First World War), and its roof story is dominated by the material-mix and climate-pace forces rather than storm turnover. Severe hail is comparatively rare, hurricanes reach the region weakened and infrequently, and so most Northeastern roofs die of old age. That makes the region the best place in the country to observe what an asphalt roof does when nothing dramatic happens to it: granule loss, curling, brittleness, and sealant failure on a 20-to-30-year schedule, with ice dams along the eaves as the signature regional injury.

The material legacy runs deep. Slate roofs from the late 1800s and early 1900s still serve on rowhouses and farmhouses across New York, Pennsylvania, and New England; standing-seam metal has a long regional history; and cedar persists on coastal architecture. On those roofs the working clock is not the surface but the fasteners, flashings, and the availability of qualified repair trades. A homeowner who inherits a slate roof inherits a roof that may outlive them — if they can find someone competent to maintain it. The practical consequence: in the Northeast, "the roof is 80 years old" can be a statement of quality rather than a crisis, and replacing sound slate with asphalt is usually a value-destroying panic move.

For the asphalt majority, the Cohort Clock applies with unusual fidelity here because storms rarely reset it. Roof ages on a Northeastern street track the street's build-and-renovation history more closely than anywhere else in the imagery record: re-roofing arrives in the neighborhood-wide clusters the cohort logic predicts, without the storm-driven patchwork common further south. A qualitative pattern rather than a statistic, but it means a Northeastern homeowner can lean harder on permits and neighbors, plus build-era arithmetic, when estimating a roof's age.

• If you live here, check: ice-dam history along eaves and valleys (interior ceiling stains near exterior walls are the tell), chimney and step flashing on older masonry, and ventilation in attics that have been insulated and air-sealed over the decades without rethinking airflow.
• If you have slate, tile, or old metal: hire material-specific expertise for assessment, and treat any contractor who quotes a same-week asphalt tear-off of a serviceable slate roof as a red flag.
• If you are buying: the region's multi-roof histories make documentation valuable. Ask specifically when the roof was last replaced and with what, not just "how is the roof?"

A Midwest case study: the postwar house on its fourth roof

Take a 1958 brick ranch in a first-ring suburb of Cleveland, a stand-in for the enormous postwar cohort that surrounds every Great Lakes city. Its original roof went on when Eisenhower was president and lasted into the late 1970s. The second was probably an overlay (new shingles nailed over the old ones, a practice common then and still permitted by many codes at one extra layer), and it wore out faster for it, coming off in the tear-off replacement of the mid-1990s. Roof three carried the house past 2010, and if the owners followed the cohort's habits, roof four is the architectural-shingle roof up there now, somewhere in its teens. Four roofs, three owners, one house: that is the normal biography of the region's postwar stock, and the Midwest's roof-age picture only makes sense read this way.

What complicates the biography is where the house sits. Ohio, Illinois, Michigan, and Wisconsin all carry older-than-average housing through real winters, with the freeze-thaw cycling and ice-dam exposure that entails, while the region's western and southern edges shade into hail country; the corridor from Kansas City through the Ohio Valley logs damaging-hail reports every spring. Where hail reaches (eastern Nebraska and Kansas, Iowa, much of Missouri and Minnesota, and episodically the whole region in a derecho year), streets show the patchwork: storm-replaced roofs a few years old beside originals from the build cohort, and a hail swath's footprint can appear in imagery as a band of newer roofs cutting across otherwise cohort-aged neighborhoods. Where storms do not reach, roofs age out on schedule the way they do in the Northeast. It is the most mixed roof-age picture in the country, and the reason this report keeps insisting that house age and roof age are different numbers.

For the owner of that Cleveland ranch, the useful moves fall out of the story. The single most valuable fact an inspector can establish on a postwar house of unknown history is the layer count, because an overlay shortens the current roof's life, hides the deck, and complicates every future job. The calendar matters too: the region's short roofing season concentrates work between late spring and early fall, so replacement bids sought in February are often better-priced and better-staffed than bids sought in June. And after any widely publicized hail event, expect door-knockers. A recorded storm near your address is a reason to get photographs of your specific roof, not a reason to sign anything on the porch.

What does roof aging look like in the Southern Plains and hail alley?

Here the storm-turnover force does not just influence the roof-age picture — it is the roof-age picture. NOAA's National Severe Storms Laboratory describes the area where Colorado, Nebraska, and Wyoming meet as "hail alley," with the most hail days per year in the country (roughly seven to nine a year in the core), and Texas, Kansas, and Oklahoma consistently log large numbers of damaging-hail reports each spring. Across this belt, from the Front Range through the Southern Plains, a large share of roofs never reach wear-out age at all: they are retired by claim, on the weather's schedule, sometimes more than once a decade for an unlucky address.

That inverts the normal relationship between house age and roof age. Texas has some of the youngest housing in the country, and its roofs run younger still; a 40-year-old Dallas-area house on its fifth roof is unremarkable. Hail-belt streets show the strongest patchwork anywhere in the imagery this report draws on: adjacent, identical houses carrying roofs a decade or more apart in age, because one storm's swath ended mid-block, or one owner filed a claim and a neighbor did not. The Cohort Clock still identifies which homes carry boom-era originals, but in this region it must always be read alongside the local storm history.

The insurance dimension is at its sharpest here, and it increasingly shapes replacement choices. The Insurance Institute for Business & Home Safety's research on hail performance underlies the impact-resistance ratings that many Plains and Front Range carriers now reward, and separate wind/hail deductibles calculated as a percentage of dwelling coverage (often 1 to 2 percent or more) have become standard in much of the belt. The practical read: in hail alley, your roof's age matters less than its rating, its documentation, and your deductible math. When a storm does come, the claim process has its own rules and deadlines; our storm damage insurance claims guide walks through them step by step.

What does roof aging look like along the Gulf and Southeast coast?

From Texas's coastal bend around Florida and up through the Carolinas, the housing stock is young and getting younger; this is where America has built most aggressively for three decades. And yet it is the region where roof age carries the most consequence, because it is where insurers scrutinize roofs hardest. Hurricane wind is the headline stressor, but the everyday ones matter too: relentless UV, heat, humidity that feeds algae and rot, and salt air near the water. Asphalt roofs here commonly age out faster than the national ranges suggest, and carriers know it.

The storm-turnover force operates differently here than in hail alley: instead of frequent local resets, the coast takes rare regional ones. A major hurricane landfall replaces an enormous number of roofs across its path within a year or two, stamping a visible age-cohort onto whole counties. That signature is readable from above in coastal metros: broad bands of same-vintage roofs tracking past storm paths, overlaid on the build cohorts. Between storms, the region's roof stock ages together toward the next test. That dynamic is why Florida's building codes since Hurricane Andrew have pushed roof attachment, secondary water barriers, and edge detailing so hard, and why the IBHS FORTIFIED standard, which specifies a sealed roof deck and enhanced attachment, has its deepest adoption along the Gulf, including Alabama's notable incentive programs.

Insurance is the region's second weather system. Percentage hurricane or named-storm deductibles are standard; carriers have non-renewed aggressively in the highest-risk zones (the December 2024 Senate Budget Committee staff report found the steepest non-renewal increases concentrated in coastal states); and roof age is a live underwriting variable at renewal. Florida has pushed back with rules that, among other things, keep an insurer from declining or dropping a policy over nothing but the roof's age while that roof is under 15, and that give owners of older roofs an inspection-based rebuttal when a report documents useful life remaining. The practical upshot for a coastal homeowner: your roof's documented age and condition are financial assets. Keep the inspection reports, the wind-mitigation forms, and the replacement invoices as carefully as you keep the deed.

• If you live here, check: roof-to-wall attachment and edge condition (the failure points wind actually exploits), the age and type of your underlayment, and whether your state or carrier offers wind-mitigation inspection credits.
• At replacement, price the resilience upgrades (sealed deck, ring-shank nails, enhanced edge metal, FORTIFIED certification where available) while the roof is open; they are cheap mid-job and impossible after.
• Re-shop your insurance after any roof replacement with documentation in hand; a new, code-current roof is one of the few levers that reliably moves coastal premiums.

The Southwest and Mountain West, in one table

What does roof aging look like in the Pacific Northwest and on the West Coast?

The West Coast's housing ages are the most mixed in the country (San Francisco and Portland cores that predate the war, Los Angeles's vast midcentury carpet, and young suburban rings everywhere), and its roof-age picture is the most maintenance-dependent. In the marine Northwest, the dominant stressor is not violence but persistence: eight months of moisture, moss and lichen colonizing north slopes, and conifer debris damming valleys and gutters. A well-maintained architectural shingle roof in Seattle can genuinely reach the long end of its range; the identical roof two doors down, under an untrimmed fir and never cleaned, can fail a decade sooner. Nowhere else does owner behavior move the roof clock this much.

California splits into its own stories. Coastal metros age gently, with mild temperatures and modest storm exposure, so the material-mix force dominates: old stock carries everything from original tile to layered asphalt histories, and roof ages track renovation economics more than weather. Inland and upslope, wildfire is the defining roof issue, both physically and financially. Class A roofing matters, ember-driven ignition through vents and debris matters more, and insurance availability in the wildland-urban interface has tightened dramatically; the same Senate Budget Committee report that documented coastal non-renewals found California's fire-exposed counties among the sharpest risers. A roof in the Sierra foothills is now underwritten as part of a home's total fire posture.

The Northwest's moss line is among the easier roof signals to read from above (the north-slope darkening and texture change that precedes real trouble), and West Coast metros otherwise show the cohort-tracking pattern of mild climates: roof ages following build and renovation waves rather than storm paths. Qualitative observations, as throughout this report. The homeowner translation is simple. In this region the inspection-and-maintenance habit is worth more than in any other, because the difference between an 18-year roof and a 28-year roof here is mostly decided by moss removal and debris clearing, plus a habit of prompt small repairs. Our roof inspection guide covers what a real condition assessment includes.

• If you live in the marine Northwest, check: moss on north and shaded slopes (treat early: established moss lifts shingle edges), valley and gutter debris every fall, and flashing at the region's ubiquitous roof-plane intersections.
• In California fire country, check: your roof's fire classification, ember-resistant venting, and defensible-space compliance. Your insurer already has.
• Everywhere on the coast: verify what is actually on the roof before assuming; the region's renovation-heavy history means listings and memories are often wrong about both material and age.

Storms are rewriting the replacement schedule

The best national evidence for the storm effect comes with an important caveat about its own future. NOAA's National Centers for Environmental Information counted a record 28 billion-dollar weather and climate disasters in the United States in 2023 and 27 in 2024. Those are the final years in that dataset: NOAA retired the Billion-Dollar Disasters series in May 2025, and the data remain archived through 2024 but are no longer updated. In both of those final years, severe storms (the category that includes hail and damaging straight-line wind, the two forces that end the most roofs) were the largest group by far, at 19 events in 2023 and 17 in 2024. Hurricanes cost more per event; severe storms happen an order of magnitude more often, and their damage is disproportionately roof damage.

Beneath the billion-dollar headline events sits the steadier drumbeat: NOAA's National Severe Storms Laboratory identifies hail alley, where Colorado, Nebraska, and Wyoming meet, as taking the most hail days per year in the country, and Texas, Kansas, and Oklahoma log large numbers of damaging-hail reports every spring, headline year or not. The Insurance Information Institute's claims data completes the picture from the payout side: wind and hail are consistently the most frequent cause of homeowners insurance claims. Put together, a substantial share of American roof replacement is scheduled by weather rather than wear. That is why this report treats storm turnover as a force that rewrites the cohort arithmetic rather than a footnote to it.

For the national roof-age picture, storm-driven turnover produces the two profiles the regional chapters described. In storm-exposed regions: younger average roof ages, patchwork streets, and a roof stock whose renewal is lumpy, arriving in post-event surges that strain contractor capacity and invite fraud. In milder regions: a steadier drift toward older roofs, with renewal following the cohort waves. Neither profile is "better." A young storm-replaced roof may carry hidden damage from the event after the one that replaced it; an old mild-climate roof may be quietly past its margins. In both cases the owner's move is the same: know the roof's actual age and condition, not its region's average.

What does an aging roof stock mean for insurance?

Insurance is where the aging of the roof stock stops being an abstraction and shows up in a homeowner's mailbox. Carriers have responded to years of rising roof losses by underwriting roof age directly, and for most owners of older roofs, the first consequence of the national picture arrives as a changed renewal. The table below summarizes the practices that have become common; our roof age and insurance guide explains each in detail, including how to respond to them.

The first symptom of an aging roof today is more often a renewal letter than a leak.

The broader pressure is documented at the national level. A December 2024 staff report from the U.S. Senate Budget Committee, an investigation into climate-driven insurance market stress based on non-renewal data collected from major carriers, found that homeowners non-renewal rates rose in most states between 2018 and 2023, with the steepest increases concentrated in storm- and fire-exposed markets along the Gulf, the southeastern coast, and in California. The report's subject is climate risk broadly rather than roofs specifically; roof age and condition are one underwriting lever inside that larger story, and this report draws no stronger link than that. Harvard's Joint Center for Housing Studies, meanwhile, has flagged rapidly rising insurance premiums as a growing share of the cost of owning a home in its recent State of the Nation's Housing reports. That cost pressure lands hardest on owners of older homes, who skew toward fixed incomes.

Some states have pushed back. Florida now stops carriers from using a roof's age as the sole basis for refusing a policy before the roof turns 15, and it lets the owner of an older roof answer an age objection with an inspection that documents remaining useful life. Other states have adjusted what carriers may require or exclude. But the direction across markets is consistent, and homeowners should plan for it: the era when a roof's age was private information between you and your gutters is over. Carriers increasingly know your roof's approximate age whether you tell them or not (roof imagery has seen to that), and the winning posture is to know it more precisely than they do, with documentation.

Put the halves of this report together and the insurance chapter almost writes itself: the country's roofs are, on average, aging into exactly the years carriers scrutinize hardest, at the same time as the institutions paying for roof failures have both the motive and the imagery to check. That collision, an aging stock meeting age-based underwriting, is the single most practical reason a homeowner should care about the national picture. It converts "my roof is fine, it's not leaking" from a reasonable position into an expensive one.

How can better roof underwriting help prevent property damage?

Roof underwriting is often described as a gate: a carrier accepts a property, changes the terms, asks for proof, or declines it. That is only the decision at the end of the process. The more useful work happens earlier, when consistent roof information helps someone notice a maintenance opportunity, document a change, plan a replacement, or send the right person to verify one unresolved condition. Better underwriting does not prevent hail or stop shingles from aging. It can reduce avoidable damage by making the next sensible action clearer while there is still time to take it.

The sequence is straightforward: identify the right roof; assemble its age evidence, material, geometry, maintenance history, visible condition, and recorded weather context; separate known facts from estimates; flag the questions that matter; verify the questions that require a physical look; and update the record after maintenance or replacement. A model is useful only to the extent that it improves that sequence. A single unexplained score may be fast, but it gives a homeowner nothing to act on and gives a professional little basis for checking whether the result is wrong.

Damage prevention is where those uses overlap. The Insurance Institute for Business & Home Safety has documented how roof design, installation, materials, maintenance, and sealed roof decks can change performance in wind, wind-driven rain, hail, wildfire, ice, and snow. Underwriting information becomes constructive when it helps direct attention toward measures like those: cleaning drainage before water backs up, correcting a loose edge or flashing detail, documenting impact-resistant materials, confirming roof-mounted equipment attachment, or planning a replacement around a stronger specification. The evidence does not perform the work. It helps the right work happen earlier.

Residential and commercial roofs need related but different models. A residential model can often organize one primary steep-slope system around age, material, replacement history, visible signals, and household decision timing. A commercial model has to track roof sections, membrane and attachment systems, drainage, rooftop equipment, repairs, warranties, inspection logs, tenant or business interruption consequences, and capital plans. RoofPredict is building both. The common principle is an evidence trail that a human can review, correct, and use—not a black-box verdict presented as certainty.

For a homeowner, the takeaway is deliberately simple. You do not need to become an underwriter. Keep the roof age and work history, preserve dated photographs and inspection reports, understand what your policy asks about the roof, and address small, verified maintenance items before they become larger ones. A well-organized roof record gives an insurer, lender, buyer, or roofer a better starting point—but its first and most valuable user is you.

What our own corpus shows

Everything to this point comes from public data. This chapter is the part of the report that comes from us. To train and validate the models that date roofs across image captures, RoofPredict keeps a sealed catalog of roof imagery: 14,050,008 images covering 2,071,744 individual U.S. homes across multiple metropolitan areas and multiple years, gathered in two broad capture waves (751,121 homes in the newer wave, 1,320,623 in the earlier one). It was assembled to teach a model what a re-roofed house looks like from above, not to survey the country, and that distinction governs everything we are willing to say from it.

What makes it useful for a report on roof aging is depth rather than breadth. 710,381 of those homes have been photographed at two or more distinct points in time, 657,876 at three or more, and 351,306 at eight or more separate dates. That last group is the interesting one: a roof seen eight times across a span of years is not a snapshot, it is a time series, and a time series is the only way to watch a roof surface change without standing on it. 6.17 million of the images carry a capture date precise to the day, which is what turns "sometime between these two passes" into a window you can actually date, and 1.067 million same-capture multiview groups show the same roof from several angles at the same moment, which is how a shadow, a tarp, or a wet slope gets told apart from new shingles.

Three patterns show up in that imagery, and all three corroborate the public data above rather than adding to it. First, re-roofing clusters: on streets built in a single cycle, replacements arrive in bunches rather than one house at a time, which is the Cohort Clock made visible from 1,000 feet. Second, storm-path turnover: in hail-belt metros, a storm's track can be read months later as a band of fresh roof surfaces cutting across a subdivision while streets a mile away stay unchanged. Third, divergence: in those same metros, roofs turn over on a schedule with little relation to when the houses were built, which is exactly why house age is a poor proxy for roof age across the storm-exposed half of the country.

Now the harder half, because a corpus this size invites overclaiming. It is not a sample of American housing and no honest national percentage can be computed from it. It was assembled where imagery is dense and where our models needed training data, so it over-represents the metros we work in and under-represents everywhere else; capture cadence differs from market to market, so a home with eight dated captures and a home with two are not equally observable; and imagery reads surfaces, so a partial re-roof can present as a full replacement, and nothing beneath the shingles appears at all. Those are properties of the data, not gaps we expect to close.

How that imagery becomes a dated estimate for one address, in what order the evidence is weighed, and where the method fails is set out on our methodology page, which carries the same corpus counts. The short version is unchanged by scale: repeat captures bracket a replacement into a range of years, a permit or an invoice beats any photograph, and someone on the roof beats both.

Replacement demand through 2030

Run the wave arithmetic forward and the rest of this decade comes into focus. The late-1990s-to-2006 cohort (the largest since the early 1970s, concentrated in the Sun Belt) has its original roofs aged roughly 20 to 28 in 2026, and each year through about 2030 pushes more of them past the typical replacement point. Stacked on top: the 1970s cohort's third-roof echo, ordinary mid-cycle replacements from the 1980s and 1990s cohorts, and whatever the weather adds. Harvard's Joint Center for Housing Studies' remodeling research consistently ranks roofing among the largest categories of homeowner replacement spending, and the Census Bureau's American Housing Survey regularly finds roofing among the most commonly reported major projects; an aging stock keeps it there. Nothing in the public data suggests replacement demand slackening before the thin 2008–2012 cohort creates its valley in the early 2030s.

This report does not forecast prices, and you should distrust anyone who forecasts them confidently: labor markets, material costs, tariffs, and storm seasons all move faster than housing demographics. What the demographics do support is a statement about leverage. A homeowner who prices a replacement while the roof is sound chooses the timing, the contractor, and the scope, while one who buys under a leak, a storm surge in demand, or a non-renewal deadline inherits whatever the market offers that week. In a decade of structurally strong demand, the gap between those two positions widens. What replacement actually costs, what drives the price, and how to read bids is the subject of our 2026 roof replacement cost guide.

One implication reaches beyond individual budgets: the wave hits hardest where houses are newest. Texas, Florida, Arizona, Georgia, Nevada, and the Carolinas absorbed the 2000s boom, so they face the deepest original-roof cohorts, under the heaviest storm exposure and the tightest insurance conditions in the country. A reader in Dallas or Tampa should hear this chapter as: your subdivision's roofs are coming due together, your insurers know it, and the neighbors who move early will have the easiest time of it. A reader in Rochester or Cleveland faces the same logic at lower intensity and on an older cohort's schedule.

The Three-Entry Roof Ledger

A national report earns its keep only if it changes what an individual owner does. This report's recommendation compresses to three entries (an age, a budget, and a file), which together we call the Three-Entry Roof Ledger. It is deliberately modest: no entry requires spending more than a few hundred dollars, most owners can complete the first in under an hour, and the whole thing exists so that the events this report describes — a hail season, a renewal letter, a wave of neighborhood re-roofing — arrive as items you check against a plan rather than emergencies.

Entry one is the keystone, and the methods for establishing it are covered stepwise in the age-detective section of the roof lifespan guide linked earlier. Entry two turns the national demand picture into a private number, and the date is a planning marker rather than a prediction: it names the year after which you stop being surprised. Entry three is the one owners skip and regret. In an era of imagery-based underwriting and percentage deductibles, dated photographs and a professional inspection report are the difference between asserting your roof's condition and proving it. Our roof inspection guide, linked above, covers what a legitimate inspection includes and what it should cost; for a middle-aged asphalt roof, a professional look every two to three years, plus after any major storm, is a common-sense cadence, moving to annual as the roof enters its final expected years.

1. This month: establish your roof's age from a real source, and take dated photos of every slope from the ground.
2. This quarter: if the roof is past 15 (or past 10 in the high-UV Southwest), get a professional inspection and file the report; then check how your policy treats roof age: payment schedule, deductibles, exclusions.
3. This year: if the Cohort Clock puts you in the replacement window, gather two or three bids on your own schedule, and start the budget entry, even if the answer is "not yet, re-check in two years."

What this report does not claim

A report that wants to be cited has to be honest about its edges, so here they are, plainly. First: there is no official count of U.S. roof ages, and we have not produced one. The housing-age figures above are Census-based; every roof-age statement is an inference from house ages, replacement-cycle arithmetic, and qualitative observation of imagery, and is labeled as such. Second: our imagery corpus is counted precisely, but those counts describe the catalog rather than the country. It covers specific metros rather than every address, it is a training and validation collection rather than a probability sample, its strengths are ranges and patterns rather than national statistics, and it reads roof surfaces; it cannot see underlayment, decking, ventilation, or flashing, which only an on-site inspection can assess.

Third: the regional and state chapters describe tendencies, not your house. A street, a subdivision, or a single storm path can invert every generalization in this report, which is why each chapter ends by pointing you back to your own roof's records. Fourth: this report does not forecast roof prices, insurance premiums, or storm seasons, and it does not claim that rising non-renewals are caused by roof age specifically; the Senate Budget Committee report we cite is about climate-driven market stress broadly, and roof condition is one factor within it. Fifth: nothing here is a substitute for a professional inspection, and no number in this report tells you whether your particular roof needs work this year. The report's claims are deliberately narrower: the stock is aging, the waves are real, the insurance response is documented, and the homeowner's best move is knowledge held in advance.

Arithmetic examples such as the Cohort Clock and replacement-wave timeline are labeled as illustrations. Approximate survey figures are presented as ranges rather than false precision. Citations should keep those labels and ranges attached to the figures so readers can distinguish a measured count from an informed estimate.

Cite this report

Reporters, bloggers, and answer engines quote what is easy to quote, and hedged prose rarely survives the trip. So here are this report's central findings written as self-contained sentences, each carrying its attribution and its underlying source. Lift them verbatim; that way the sourcing travels with the number instead of falling off along the way.

• According to RoofPredict's 2026 National Roof Age Report, the median owner-occupied U.S. home was 42 years old in 2024, up from 31 years in 2005 (NAHB analysis of Census American Community Survey data).
• According to RoofPredict's 2026 National Roof Age Report, roughly half of America's owner-occupied homes were built before 1980, while only about an eighth have been built since 2010 (NAHB analysis of Census American Community Survey data).
• According to RoofPredict's 2026 National Roof Age Report, U.S. housing starts climbed past two million in 2005 and collapsed to roughly 550,000 by 2009, so the original roofs on that boom cohort are reaching replacement age together through about 2030 (U.S. Census Bureau, New Residential Construction).
• According to RoofPredict's 2026 National Roof Age Report, NOAA counted a record 28 billion-dollar weather and climate disasters in 2023 and 27 in 2024, with severe storms the largest category in both years; NOAA retired that series in May 2025 (NOAA National Centers for Environmental Information).
• RoofPredict's 2026 National Roof Age Report draws on a proprietary roof imagery corpus of 14,050,008 roof images covering 2,071,744 U.S. homes, 710,381 of which have been photographed at two or more distinct points in time (RoofPredict roof imagery corpus, as of August 2026).
• According to RoofPredict's 2026 National Roof Age Report, no U.S. agency counts roof ages nationally, so every published national roof-age figure, including any inferred from this report, is an estimate rather than a count.

A full citation reads: RoofPredict Research Team, "The 2026 National Roof Age Report: How Old Are America's Roofs?", RoofPredict, updated August 20, 2026, https://roofpredict.com/resources/2026-national-roof-age-report. If you quote the corpus figures, please point readers to https://roofpredict.com/resources/methodology as well, since that page states plainly what the imagery can and cannot establish. Corrections are welcome and will be made in place.

Methods, sources, and update cadence

The evidence in this report sits in three tiers, used in a strict order. Tier one is named public data: the Census Bureau's American Community Survey (including the year-structure-built tabulations in table B25034) and New Residential Construction series, NAHB's analyses of that Census data, Harvard's Joint Center for Housing Studies reports, NOAA's NCEI disaster series and NSSL hail climatology, the Insurance Information Institute's claims data, and the Senate Budget Committee's December 2024 staff report on insurance market stress. Every specific number in this report traces to that tier. Tier two is labeled arithmetic (the Cohort Clock and the wave timeline), which applies standard replacement-cycle assumptions to tier-one data and claims nothing more than illustration. Tier three is RoofPredict's own roof imagery corpus: 14,050,008 images over 2,071,744 U.S. homes, sized in the corpus chapter above and on our methodology page. Those counts are measured from the catalog rather than estimated, and they are the only first-party figures in this report; what we read out of the imagery itself stays strictly qualitative (cohort re-roofing clusters, storm-path bands), because a collection built where imagery is dense cannot support a national percentage.

How that imagery tier works, including its accuracy characteristics and known failure modes, is documented on our methodology page. The short version: comparing historical imagery passes brackets a roof's last replacement into a range of years; the estimate is only as good as the imagery coverage for an address; and surface reading has inherent limits that on-site inspection does not. The report uses ranges and "illustrative" labels wherever the evidence does not support a more precise statement.

This is a living document. We expect to revise it when the Census Bureau releases new ACS year-built tabulations, when NAHB updates its housing-stock age analyses, and as the post-NOAA-series landscape for disaster-loss data settles; the updated date at the top of the page reflects the last substantive revision. Corrections are welcome and will be made visibly. If the numbers here have done their job, the next step is local: find out where your own roof sits on the clocks this report describes, and write the three ledger entries that turn the answer into a plan.

Frequently asked questions

How old is the average roof in the United States?

No public agency measures roof age nationally, so every figure you see quoted, including any you might infer from this report, is somebody's estimate rather than a count. What the public record does support: NAHB's analysis of Census data puts the median owner-occupied home at roughly 42 years old, asphalt roofs are typically replaced somewhere inside a 15-to-30-year window, and the enormous 2000s building cohort is arriving at replacement age all at once. Put those together and most American roofs sit somewhere in their first two decades of service. The average gets pulled younger across the hail belt and the storm-exposed coasts, where weather retires roofs long before wear does, and older across the Northeast, where it rarely intervenes. The number that matters to you is not the national one anyway. It is the install date on your own roof.

Which states have the oldest homes?

NAHB's state-level analysis of Census data puts New York oldest among the 50 states, with a median owner-occupied home above 60 years, and Massachusetts and Rhode Island close behind; the District of Columbia is older still. The youngest stock sits in the Sun Belt and Mountain West, led by Nevada and Texas.

Which regions have the oldest roofs?

The oldest working roofs concentrate in the Northeast. New York and New England have the country's oldest houses, storms rarely retire roofs early there, and long-lived slate and metal are part of the stock, so roof ages track the old build cohorts. The youngest roofs cluster where weather does the retiring: the hail-prone Southern Plains and Front Range, and the hurricane-exposed Gulf and Southeast coasts. One caution: house-age data is Census-based, while the roof-age reading is an inference, since no agency counts roof ages directly.

Does an older house mean an older roof?

Not reliably. Roofs are replaced several times over a house's life, and storms often reset the clock; a 40-year-old Dallas house may have a 4-year-old roof. Find the roof's own record, not the house's.

What is the Cohort Clock?

This report's shorthand for aging a roof by arithmetic: count forward from the build year in 20-to-25-year asphalt cycles and see where 2026 falls. A starting guess, not a measurement.

Why are homes built in the early 2000s being singled out?

Because so many were built at once. Census construction statistics show housing starts passing two million a year in 2005 and collapsing to roughly 550,000 by 2009. The original roofs on that boom cohort are now about 20 to 28 years old, so a very large group of homeowners is reaching the same decision within the same few years.

Was 2023 or 2024 a record year for billion-dollar disasters?

2023 holds the record at 28 billion-dollar weather and climate disasters; 2024 counted 27, per NOAA's National Centers for Environmental Information. Those are the final entries in the series, because NOAA retired the Billion-Dollar Disasters product in May 2025 and the data remain archived through 2024 without further updates. Severe storms, the category covering hail and damaging wind, led both years, with 19 events and then 17.

Is the aging housing stock why my insurance went up?

It is one factor among several. Harvard's Joint Center for Housing Studies and a December 2024 Senate Budget Committee staff report both document rising premiums and non-renewals, driven by storm losses, rebuilding costs, and reinsurance prices as well as roof condition. What you can influence is your own roof's documented age and condition, which many carriers now price directly.

Will roof replacement get more expensive because of this?

We do not forecast prices. What can be said: a large cohort of roofs is reaching replacement age together, storm-driven replacements add demand on top, and roofing is already among the most common major home projects. That argues for pricing a replacement while your roof is still sound, on your own schedule, rather than under a deadline set by a leak or a non-renewal letter.

How many roofs are replaced in the U.S. each year?

No agency publishes an official count, and the reason is structural: re-roofing permits are issued locally, in thousands of separate jurisdictions, on inconsistent rules, and a meaningful share of work never generates a permit record at all. What the public sources do establish is scale. Harvard's Joint Center for Housing Studies consistently ranks roofing among the largest categories of homeowner replacement spending, and the Census Bureau's American Housing Survey regularly finds roofing among the most commonly reported major projects. With tens of millions of homes past age 20 and asphalt cycles running 15 to 30 years, annual replacements plausibly run into the millions. Treat that as an inference from the size of the stock rather than a counted statistic, and expect big storm years to swing it hard.

What did RoofPredict analyze for this report, and how far can that kind of estimate be trusted?

Our contribution is a first-party roof imagery corpus, built to train and validate RoofPredict's roof-dating models: 14,050,008 images covering 2,071,744 U.S. homes across multiple metros, 710,381 of them photographed at two or more distinct points in time. Comparing captures of the same rooftop brackets a last replacement into a window of a few years rather than pinpointing a date. Every national statistic here comes from somewhere else: the Census Bureau, NAHB, NOAA, Harvard's Joint Center for Housing Studies, and the other public sources listed below. The corpus shows up only qualitatively, in observations like cohort re-roofing clusters and storm-path turnover, and never as a percentage, because a catalog assembled where imagery is dense is not a national sample. The limits are worth stating plainly. Imagery cannot see underlayment, decking, or ventilation; a partial re-roof can read like a full one from above; and capture frequency varies by metro. A permit record or an invoice sharpens any such estimate into an actual date.

How big is RoofPredict's roof imagery dataset?

As of August 2026 the corpus holds 14,050,008 roof images covering 2,071,744 U.S. homes, gathered in two capture waves of 751,121 and 1,320,623 homes. 710,381 of those homes have been photographed at two or more distinct points in time, 657,876 at three or more, and 351,306 at eight or more, which is what makes it a roof time series rather than a pile of pictures; 6.17 million images carry a capture date precise to the day, and 1.067 million same-capture multiview groups show one roof from several angles at one moment. It exists to train and validate roof-dating models across multiple U.S. metros, so its counts describe the catalog and not the national housing stock.

Does the Cohort Clock work for tile, metal, or slate roofs?

Not directly, since those materials run far longer cycles and a build year tells you correspondingly less. Tile carries a catch worth knowing: the underlayment beneath it ages on roughly an asphalt schedule, and that layer usually fails first.

Does this report mean I need a new roof?

No. Averages describe the stock, not your house, and plenty of 25-year-old roofs are sound. Learn your roof's real age, then inspect on a schedule that fits it.

Sources

• NAHB Eye on Housing series on the age of the U.S. housing stock, analyzing Census ACS data
• U.S. Census Bureau, 2024 American Community Survey, table B25034: Year Structure Built
• New Residential Construction historical data, U.S. Census Bureau
• American Housing Survey, U.S. Census Bureau
• Joint Center for Housing Studies of Harvard University, The State of the Nation's Housing 2025
• U.S. Billion-Dollar Weather and Climate Disasters archive, NOAA National Centers for Environmental Information (series retired May 2025; data through 2024)
• NOAA National Severe Storms Laboratory, Severe Weather 101: Hail
• Insurance Institute for Business & Home Safety, Roofing Roadmaps
• RICOWI and Insurance Institute for Business & Home Safety, Best Practices Guide for Roofing
• Insurance Information Institute, "Facts + Statistics: Homeowners and Renters Insurance" (locate by title on iii.org)
• RoofPredict roof imagery corpus (proprietary), described at /resources/methodology

Figures reflect general industry guidance and public data, not a prediction about any specific roof. See our methodology and editorial standards.