A rancher checking fence lines in eastern Montana has no cell signal for 40 miles in any direction. A pipeline inspector in West Texas watches her tablet spin while it tries to upload a safety report. These are the customers T-Mobile says it can finally reach. The carrier is packaging its 5G network with SpaceX’s Starlink satellite technology into a single managed service for business customers, making it the first major U.S. carrier to sell satellite and cellular connectivity under one enterprise plan.
The promise is straightforward: when terrestrial towers disappear, a Starlink satellite overhead picks up the connection, all on one bill and one support contract. But independent research published in spring 2025 shows the satellite side of that equation is still in its early stages, and businesses considering the bundle need a clear picture of what works today versus what is still on the roadmap.
What independent testing actually shows
The most rigorous public evidence about Starlink’s direct-to-cell performance comes not from T-Mobile or SpaceX but from an independent technical study. A research preprint hosted on arXiv, the Cornell University-affiliated repository for scientific papers, presents crowdsourced measurement data collected across the United States between October 2024 and April 2025. Real users carrying standard smartphones contributed the readings, giving the study a ground-level view of how the network behaves outside a lab.
The headline finding: during that seven-month window, the Starlink direct-to-cell network supported SMS only. No voice calls. No mobile data sessions. For a business hoping to run cloud applications, stream video from a job site, or maintain a VoIP call when 5G drops out, that is a hard constraint. Through at least April 2025, the satellite link functioned as a text-messaging lifeline, not a broadband replacement.
Signal quality told a similar story. The researchers found that satellite signals were consistently weaker than those from conventional cell towers. That gap is partly physics: a satellite hundreds of kilometers overhead simply cannot match the signal strength of a tower a few kilometers away. But it matters for any plan that advertises seamless failover. If the satellite leg delivers noticeably weaker connectivity, the transition from 5G to satellite may feel less like a backup and more like a downgrade, especially for data-intensive tasks such as fleet tracking or remote equipment monitoring.
Geographic patterns emerged as well. Urban and suburban areas with dense tower coverage produced few satellite interactions, while rural zones with sparse infrastructure generated more direct-to-cell events. Even in those rural areas, though, the service stayed within its SMS-only boundary. Connection setup times for satellite-routed texts were also longer than for standard cellular messages, reflecting the propagation delay inherent in bouncing signals through orbit. For routine check-ins, the delay is manageable. For time-critical safety alerts, it is a factor worth building into operational planning.
Regulatory groundwork already in place
One piece of the puzzle is more settled than the article’s technical findings might suggest. In late 2024, the FCC granted T-Mobile and SpaceX authorization under its supplemental coverage from space (SCS) framework, allowing Starlink satellites to use T-Mobile’s mid-band spectrum to connect directly to standard cell phones. That approval was a regulatory first and removed a significant barrier to commercializing the service.
The SCS authorization means the bundled plan is not operating in a regulatory gray area. T-Mobile has spectrum rights for the satellite link, and the FCC has formally blessed the arrangement. What the authorization does not guarantee is performance. Spectrum access and network capability are different things, and the independent measurements from spring 2025 show the capability side is still catching up to the regulatory green light.
What businesses still do not know
Several critical details remain unconfirmed as of June 2026. T-Mobile has not publicly disclosed final pricing for the bundled enterprise plan, specific rollout dates by region, or which business segments it intends to prioritize. Without those details, it is hard to judge whether the service will be priced for a 10-person ranch operation, a Fortune 500 logistics fleet, or both.
Technical specifics about the handover protocol between terrestrial 5G and the Starlink satellite layer have not been published either. Latency benchmarks beyond the SMS use case documented in the arXiv study are absent from the public record. For an enterprise IT team, the difference between a 50-millisecond and a 500-millisecond handover could determine whether a video call survives the switch or a remote sensor loses its connection.
Capacity under load is another open question. Satellite bandwidth is finite. If thousands of enterprise customers hit the direct-to-cell link simultaneously during a natural disaster or a major terrestrial outage, performance could degrade. The arXiv study did not measure behavior under heavy congestion, so there is no independent data on how the system holds up when demand spikes.
SpaceX and T-Mobile may have upgraded the direct-to-cell system after the study’s April 2025 cutoff, potentially adding voice or limited data capabilities. Any such improvements fall outside the measurement window, so current performance could differ from what the researchers recorded. Until new independent testing surfaces, the preprint remains the best public benchmark.
How this fits the competitive landscape
T-Mobile is not the only carrier chasing satellite-to-phone connectivity. AT&T has partnered with AST SpaceMobile, which launched its first commercial test satellites in 2024 and aims to deliver broadband-speed direct-to-cell service. Verizon has explored satellite partnerships of its own. The race matters because whichever carrier delivers reliable voice and data from orbit first will have a significant edge in selling to enterprises that operate beyond tower range.
T-Mobile’s advantage right now is structural: it has FCC authorization, an active beta with real user data, and a named satellite partner in SpaceX that has already deployed direct-to-cell hardware in orbit. AST SpaceMobile’s technology promises higher throughput per satellite but is earlier in its commercial timeline. For a business buyer, the practical question is not which technology is theoretically superior but which one can deliver a signed service-level agreement with measurable uptime guarantees today.
A practical framework for enterprise buyers
For IT teams evaluating the T-Mobile and Starlink bundle, the first step is concrete: catalog every connectivity task each remote site requires, from basic text alerts to real-time data uploads and voice communication. Map those needs against the SMS-only baseline documented through April 2025.
Any site whose operations depend on more than messaging should wait for independent verification of voice and data capabilities before committing to a contract. When engaging T-Mobile, ask for written service-level agreements that specify what happens when traffic shifts from 5G to satellite. Key metrics to pin down include expected latency during handover, guaranteed uptime percentages for the satellite leg, and support response times when the connection fails.
For operations where even basic text connectivity in a dead zone would be valuable, such as agricultural monitoring, remote safety check-ins, or disaster-recovery communication, the bundle could deliver real, immediate benefit. A text message that gets through when nothing else works is not a small thing.
The bottom line for enterprise decision-makers is use-case specificity. Starlink’s direct-to-cell link, as independently measured through spring 2025, is best understood as an SMS-based resilience layer that extends minimal connectivity into places where traditional networks fail entirely. That is genuinely useful for the right scenarios. It is not yet a universal fix for remote operations, and treating it as one before the data supports it would be a costly assumption.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
