Address-level peril intelligence

282.9M U.S. addresses, graded for risk.

A purpose-built deterministic peril model that grades 282.9 million U.S. addresses across 54 perils — patent-pending under USPTO 64/070,786. Address-level resolution. Validated against decades of observed events. Audit-ready intelligence for the way you actually price risk.

Request access See how it works
282.9M U.S. addresses · 54 perils · Validated methodology
Atlas / Continental US 282.9M addresses indexed
Continental United States risk gradient
ALow
BModest
CModerate
DElevated
FSevere
Atlas data feed — UTC refresh 30s · methodology validated
/v1/risks/fast47msp95 24h cache hit99.94%edge throughput12,847 req/speak 60s addresses indexed282,904,029graph nodes convective cycle14ZT+18m fresh forecast run12Z16-day fwd radar scan14:30Z2-min cadence convective outlookENHTX · OK · KS tornado watches3active wildfire incidents471.2M ac YTD air-quality index152ABQ · USG flood layerv2024.2refreshed 12d seismic shakenightly02:14Z perils modeled54all major classes SFHA addresses5,389,072addrs · 2.45% model SLA99.97%trailing 90d
How it works

Address in. Audit-ready score out.

Our deterministic peril model spans 282.9 million addresses across 54 perils, calibrated against two decades of observed severe events and a sixteen-day forward layer. Every score is timestamped, versioned, and accompanied by a reason code — patent claim 30, USPTO 64/070,786 — built to survive regulatory examination.

01 · Address

Submit any U.S. address or parcel

Single call, batch CSV, or streaming. Rooftop geocoding plus parcel-shape resolution where assessor records exist.

› 1242 Ocean Dr, Miami, FL 33139
02 · Resolve

Cross-reference the hazard catalog

Fifty-three peril layers, climate normals, and historical events. Each value is timestamped, versioned, and cached to a parcel hash.

HURRICANE FLOOD WILDFIRE SEISMIC HAIL TORNADO SURGE +18
03 · Score

Return a graded peril profile

Letter grade, percentile rank, top contributing perils, and full methodology trace — all in one sub-200ms response.

F 98th %ILE
Miami Beach parcel HURRICANE · COASTAL FLOOD · STORM SURGE
0M
U.S. addresses graded across the lower 48 plus AK and HI.
0
Distinct perils modeled — flood, fire, wind, seismic, climate-chronic.
0 yr
Of validated severe-event observations in the calibration catalog.
0ms
Cold p95 API latency on /v1/risks/fast — single-PK lookup at carrier scale.
Capabilities

Forty-six perils, one schema.

From acute hurricane risk to chronic heat-stress trajectories, every peril is normalized to a 0–100 percentile and an A–F grade with full methodology trace.

Hurricane
Coverage100% Atl. coast
Wildfire
WUI parcels44.8M
Flood (riverine)
SFHA layerLive
Storm surge
Surge zonesAll Cat 1–5
Tornado
EF tracks1950–today
Hail
Convective22-yr archive
Earthquake
Seismic hazard50-state
Heat stress
Chronic2050 horizon
See all 54 perils
Industries

Built for the balance sheet, not the demo reel.

Atlas United is engineered for the workflows that actually price and accept risk: actuarial, underwriting, accumulation management, and public-program design.

01 · Carriers

Personal & commercial lines

Underwrite admitted books with a rate-filable peril score and accumulation control at policy inception.

For carriers
02 · MGAs

Programs & specialty

Stand up new programs faster — surplus, parametric, and niche peril MGAs use Atlas as their primary risk lens.

For MGAs
03 · Reinsurance

Treaty & ILS

Portfolio rollups, PML/AAL benchmarking, and cession-quality validation for treaty pricing and ILS structuring.

For reinsurers
04 · Public sector

Federal, state, local

Mitigation grant prioritization, hazard mapping, and BRIC/HMGP application support — auditable, citable, defensible.

For public sector
05 · Research

Academia & think tanks

Citable peril datasets for peer-reviewed climate, housing, and equity research with reproducible methods.

For research
Methodology

Eight model layers. One unified grade.

Atlas United is a purpose-built deterministic peril model. Each layer is independently calibrated against decades of observed events and folded into a single audit-ready grade per address — no black-box blending, no proprietary aggregator math.

ATMOSAtmospheric & tropicalHurricane · Climatev2024.3
SURFACERiverine & coastal floodSFHA · Surgev2024.2
SEISMICEarthquake & tsunami50-state hazard2023 catalog
WILDFIREBurn & WUI exposureWUI · Embersv2024.1
CONVECTIVEHail · tornado · severe wind22-yr archive1955–today
CHRONICHeat · drought · sea-level2030 · 2050 · 2100Climate normals
HISTORICALValidated event catalog22 years · all perilsCalibration set
FORECASTForward probability layer16-day horizonSub-hour refresh
Read the full methodology
Compare

Why teams move off legacy vendors.

Legacy stacks were built when peril modeling meant licensing one vendor's black box and trusting the output. Atlas United is a transparent deterministic model with per-field methodology trace — designed for the way underwriters and regulators actually work.

Atlas United Verisk / AIR Cotality First Street
Transparent methodology ✓ Per-field Vendor docs Vendor docs ✓
Cited provenance per score ✓ Per-field Vendor docs Vendor docs Partial
Sub-200 ms address API ✓ p50 ~10 ms Batch only Available ✓
All 54 perils, one schema ✓ Fragmented Fragmented 5 perils
Federal-source citation per response ✓ Per-field No No Methodology PDF
Per-call pricing, no seat fees ✓ License License Tiered
See full comparison
"
A peril score that ships with the model version, the calibration vintage, and a cryptographic receipt. The methodology is the product.
Atlas United founder note
Get started

The peril stack your regulators already trust.

30-day evaluation with API keys, sample data for your book, and a deployment engineer assigned to your account.

Request access Read the docs
0M U.S. addresses indexed across 54 perils
0yr Validated event catalog calibrating the model
0ms Cold p50 latency on the fast endpoint