Every roofer we talk to has the same instinct: storms are getting worse. The chases are longer, the hail is bigger, the claim counts keep breaking records. We wanted to know if the instinct matched the data. So we pulled five years of NWS severe-weather warning archives and bucketed every single warning using the same tiering logic the live product ships with.
The result was sharper than we expected — and it tells a story that is subtly different from the one most people assume.
Severe storm volume is up +64% — Extreme-tier storms up +209%
Count of US NWS severe-weather warnings by roofing-weighted tier, year over year. Each year contains all three tiers — the story is how much each one is growing.
The right story: more severe storms, not stronger ones
If you squint at that chart, the first instinct is “every storm is getting more severe.” That is not actually what the data shows. Every year on that chart contains the same three tiers of warnings — Severe (yellow), Very Severe (orange), and Extreme (red). A Severe thunderstorm warning in 2021 and a Severe thunderstorm warning in 2025 are rated the same way by the same classifier. Individual storms are not getting mysteriously stronger.
What is changing is the volume of severe storms — and the mix is shifting toward the higher tiers. The total number of severe-weather warnings issued in the US climbed from 3,660 in 2021 to 6,020 in 2025, a +64% jump in four years. But the most-damaging tier grew nearly three times faster.
The Extreme tier is the real headline
Extreme-tier warnings (large/giant hail, destructive winds, tornado emergencies, IBW-Destructive-tagged events) grew +209% between 2021 and 2025 — while the Severe tier grew just +52% in the same window.
This is the structural shift hidden inside the headline number. More storms overall, yes — but the fastest-growing slice of that pie is the one that does the most damage. That is the difference between “we are just seeing more weather” and “we are seeing more weather that matters.”
Concretely: in 2021 the average roofing company in an active market had to track about 110 Extreme-tier events nationally across the full year. In 2025that number was 340. By 2026 we project 420. The companies that win each storm season are increasingly the ones equipped to mobilize on the worst-tier events fast — because that is where both the claim economics and the competitive pressure are concentrating.
The insurance industry already priced this in
This volume curve is not just a weather curve — it is also a loss curve. Insured severe-convective-storm losses in the US moved roughly in lockstep with the warning-count curve:
- 2021: ~$35B in insured SCS losses
- 2022: ~$35B
- 2023: ~$61B — first year above $60B on record at the time
- 2024: ~$70B — and the record did not hold long
- 2025: ~$70B+ — the second-costliest SCS year ever
Reinsurance carriers have already responded. Wind/hail deductibles in Texas, Colorado, Oklahoma, and Kansas have been restructured to percentage-based rather than flat in most markets. Several major carriers have pulled back on new business in the most exposed ZIP codes. Rising warning volume + a faster- growing Extreme tier = tightening claim economics is the landscape every roofing company now operates in, whether they have priced for it or not.
The 2026 projection
Based on the 5-year trajectory plus preseason guidance from Colorado State, NOAA, and AccuWeather for the 2026 severe- weather and hurricane seasons, here is what we project for the year ahead:
~11.1 million US homes and ~745 commercial buildings will fall inside at least one severe-weather warning polygon during 2026 — a new record in both categories. Total severe- weather warnings: ~6,570. Extreme-tier warnings specifically: ~420 — roughly 3.8× the 2021 baseline.
Two things to be honest about:
- This is a projection, not a forecast. Severe weather is still fundamentally a week-by-week, event-by-event phenomenon. Any single year can come in above or below trend by a meaningful margin.
- The direction is more reliable than the magnitude. Even if 2026 comes in 10% below our projection, it still represents more Extreme-tier warnings than any year on record, because the baseline has been climbing for half a decade.
What this means operationally
For a roofing company, the operational takeaway is not “buy more trucks.” It is that the weight of storm response has shifted toward the top of the severity distribution. More warnings, and disproportionately more at the severity level that actually produces claim-worthy damage, means:
- Speed compounds. The lag between a storm hitting and a crew having a lead list costs more revenue every year because the upper-tier events drive outsized claim volume.
- Out-of-state chaser density is going up. Severity-weighted claim economics attract more chasers to more markets. Local contractors who were insulated by geography are seeing out-of-state crews in neighborhoods that never used to see them.
- Documentation discipline matters more. Carriers are scrutinizing age, condition, and pre-existing wear harder every renewal cycle. Photographic and Xactimate discipline has gone from “nice to have” to table stakes for insurance work.
- Commercial is a growing tier. The commercial projection above (~745 buildings) has grown nearly 40% in five years. Contractors with commercial capability are increasingly better positioned than pure- residential shops.
How we calculated this
Every count in the chart comes from running archived NWS CAP warnings through the severity-keywords classifier that ships in the live product. The classifier reads event type, IBW tags, peak-hail and peak-wind magnitudes, and warning geometry, then assigns each warning to exactly one of the roofing-weighted tiers: Severe, Very Severe, or Extreme.
The residential- and commercial-impact counts are derived from year-end tallies of US properties falling inside the issued warning polygons, using the same parcel-level geocoding that powers the active product. Historical counts are reconciled with NOAA NCEI and SPC annual reports; 2026 is a trend projection, flagged visually in every chart it appears in.
If you want to see where the storms that drive this trend are happening right now, the live map on our Recent Damage page shows every severe-weather warning in the last 24 hours, tiered against the same model. And if you want to turn any one of those warnings into a filtered property list in about two minutes, that is literally what the product does.
The five-year curve says the opportunity cost of being slow is going up every year — and going up fastest at the tier that actually pays the bills. The companies that win 2026 will be the ones who figured that out first.