Every major US carrier claims nationwide coverage, and each claim is defensible under its own definition. Understanding what is being measured explains why maps and experience diverge.
Population covered is not area covered
Coverage is conventionally expressed as a share of the population within a predicted signal area, because networks are built where people are.
Population concentrates heavily, so covering most people requires covering a much smaller fraction of the country's land area.
A map colored across a state can therefore represent a claim about who lives there rather than about where a signal is available on any given road.
Maps are propagation models, not measurements
Coverage maps are generated by software that models signal from known tower locations across terrain and clutter data, then draws a boundary at a chosen threshold.
Models are calibrated against sampled measurements, but they cannot account for a specific building's construction, a hill between a house and a tower, or seasonal foliage.
The threshold chosen also matters: a level sufficient for a voice call is well below what a video stream needs, and maps do not always distinguish the two.
Roaming fills gaps invisibly
Carriers hold agreements letting their subscribers use other networks where they lack their own facilities, and that coverage appears on the map like any other.
Roaming service is often restricted in ways the map does not show, including data caps, lower priority and unavailability of operator features.
A phone can therefore show full bars in a claimed-covered area while behaving as though the connection were much weaker.
Federal reporting has tightened the definitions
Regulators now collect standardized coverage data from carriers and accept public speed measurements as evidence, with a challenge process for disputed areas.
That framework exists because earlier self-reported maps overstated availability, particularly in rural areas where funding decisions depended on them.
The resulting public maps are more conservative than carrier marketing maps, and they are the better reference when the question is whether service exists at all.
Testing beats mapping for an individual address
Because the variables that matter most are local, the only reliable check is running a device where it will be used, indoors, at the times of day that matter.
Return windows and trial periods exist partly for this, and prepaid service on a carrier's own network is an inexpensive way to test before committing.
Crowdsourced measurement maps help as a second opinion, since they report what devices actually recorded rather than what a model predicted.