SpaceX plans to take on T-Mobile, AT&T, and Verizon with cell gear bolted to Starlink dishes

On SpaceX's first earnings call, president Gwynne Shotwell said the company plans to build a terrestrial mobile network that will "acquire quite a few" of the customers now subscribed to T-Mobile, AT&T, and Verizon (The Verge). Until now, SpaceX has run Starlink Mobile as a partner to those carriers, not a competitor.

Shotwell said the 65 megahertz of spectrum SpaceX bought from EchoStar, the long-running satellite operator, "does have terrestrial components" and that SpaceX "definitely intend[s] to build out the terrestrial component," adding the hardware and systems for "a true mobile service" (The Verge). Musk then named the surprise: instead of "very expensive and difficult to locate, large cellular base stations," SpaceX can "deploy a large number of small stations." Those small stations are femtocells, the low-power cellular base stations the industry has sold for home and office use for years, bolted to the same mount that already holds a Starlink broadband dish.

Femtocells normally route calls and data over a wired internet connection. The SpaceX idea skips the wire and skips the tower: the dish already brings Starlink broadband into the home, and a femtocell on the same mount lets the home act as a tiny cell site for any phone in the EchoStar spectrum bands. The strategy trades spectrum scale for cell-site density on rooftops SpaceX already reaches.

That trade is the central question. The 65 MHz SpaceX acquired is a small fraction of what each of the Big 3 wireless carriers holds (The Verge). Musk hedged accordingly: "we feel reasonably confident" the rooftop-femtocell approach will scale, not certain. The mechanism is also the falsifier: if a Starlink dish is too far from most phones in a neighborhood, the cell signal stops short of a tower's reach.

SpaceX already runs a service called Starlink Mobile, a direct-to-device offering that augments carrier networks, including T-Mobile's in the US, with limited messaging and data in dead zones where cell towers do not reach (PCMag). Shotwell said that service will be "100 times better" once it pairs the EchoStar spectrum with Starlink Mobile (V2) satellites that "will start flying next year." V2 is SpaceX's next-generation direct-to-cell satellite, designed to relay ordinary phones to the network from orbit without specialized hardware.

The regulatory scaffolding is in motion. The Federal Communications Commission has already conditionally granted a SpaceX/T-Mobile waiver for supplemental coverage from space (FCC DA-25-197) and accepted the EchoStar-to-SpaceX spectrum transaction for filing in September 2025 (FCC DA-25-917). SpaceX has separately asked the FCC for approval to operate 15,000 satellites in the same MSS, or mobile satellite service, spectrum (Satellite Today), a buildout that, combined with the rooftop femtocells, would give SpaceX both an orbital and a ground layer for reaching ordinary phones.

T-Mobile, AT&T, and Verizon have not publicly responded. The earnings call transcript does not support a guess about their next move. What the earnings call did say, in plain English, is that SpaceX believes it can build a competing mobile network without the macro towers the Big 3 spent two decades erecting. Whether the rooftop-femtocell density actually covers a neighborhood, and whether customers will switch carriers to get it, are the two questions SpaceX will have to answer when the V2 satellites start flying.