US mobile download speeds surged during the first half of 2026. The combined nationwide median across AT&T, T-Mobile, and Verizon reached 334 Mbps, up about 74% from 192 Mbps in the first half of 2024, according to Ookla’s RootMetrics unit.
The increase strengthens the case for cellular connectivity at branch offices, temporary sites, and locations with limited wired options. National rankings, however, cannot show whether a carrier will deliver stable indoor service on an organization’s hardware and applications.
The H1 2026 State of the Mobile Union report draws on about 3 million tests conducted from January through June. RootMetrics covered all 50 states and 125 major metropolitan markets, logging more than 243,000 drive-miles and conducting more than 6,900 indoor tests.
The share of download tests using 5G reached 93%, up from 84% in the prior-year period.
T-Mobile wins on speed; Verizon takes the overall lead
T-Mobile recorded the highest nationwide median download speed at 392 Mbps, up from 374.5 Mbps in the second half of 2025. Verizon followed at 302 Mbps, rising from 226.5 Mbps, while AT&T increased from 193 Mbps to 228.7 Mbps.
Verizon nevertheless finished first in RootMetrics’ broader national ranking with a RootScore of 98.4. AT&T scored 98.3, and T-Mobile received 97.3.
RootMetrics evaluates performance across its nationwide sample rather than ranking carriers solely by their highest aggregate download median, according to Light Reading’s analysis. T-Mobile led on median download speed, while Verizon’s performance across the nationwide test set produced the higher overall score.
The tests used current flagship Android phones, while business deployments often rely on cellular routers, embedded modems, older tablets, and mixed-generation device fleets. Performance on those devices may differ from the published medians.
National results also cannot predict service inside a particular office, warehouse, hospital, or multistory facility. Building materials, congestion, device antennas, and nearby cell sites can all affect performance as organizations increasingly treat wireless connectivity as core infrastructure.
Fast national results still require local testing
The findings support evaluating cellular service for branch failover, field operations, temporary sites, and locations where wired connections are unavailable or slow to provision.
Using 5G as a primary connection requires closer validation. Download medians do not establish upload capacity, latency, packet loss, indoor signal quality, or performance when multiple users compete for bandwidth.
IT teams should test each location during normal and peak periods using the equipment and applications planned for deployment. Mobile speed-testing tools can establish an initial baseline, but production evaluations should also measure:
- Upload speed, latency, jitter, and packet loss under concurrent workloads
- Indoor signal and sustained throughput on the proposed hardware
- Failover, reconnection time, and carrier consistency at every site
Video meetings, virtual desktops, cloud backups, and field-data uploads may depend more on stable uplink performance and latency than maximum download speed. Broader infrastructure research likewise shows how congestion and network design can constrain applications even when headline capacity appears sufficient.
All three carriers delivered substantially faster service than two years earlier. The best enterprise option will still be the network that performs consistently at the organization’s locations, on its equipment, and under peak workloads.
Read more: Organizations weighing alternatives to terrestrial cellular service can also examine how SpaceX’s proposed Gen3 satellite network could expand future connectivity options, although regulatory approval, pricing, and availability remain unresolved.