SpaceX Designs Rooftop Cell Tower to End Dead Zones

SpaceX has finished designing a miniature cell tower that sits on a roof like a Starlink dish, and Elon Musk says it could go into production next year. The SpaceX chief executive described the device during an interview at the New York Times DealBook Summit, saying it is intended to solve the one place where satellite phone coverage fails: indoors.
"We've got a plan that enables ubiquitous access on Earth, including via cell phone," Musk said, before naming the gap in it. "If you're indoors, especially if you have a metal roof, that blocks the signal from space." The answer, he said, is a small ground station that brings the network inside the building.
A rooftop box, not a tower
Musk described the device as "a micro cell tower thing that you just put on your roof, like a Starlink dish". It would be a fully functional, low-power cellular base station rather than a simple repeater, operating on mobile spectrum SpaceX has acquired and using Starlink as the connection back to the wider network in places without cable or other infrastructure.
SpaceX president Gwynne Shotwell had already signalled the direction in August, telling an earnings call that the company would deploy small, low-power cellular base stations rather than big towers. Musk added that the rooftop units could be bundled with Starlink terminals, and that in densely populated areas the company envisages many small cell sites scattered through neighbourhoods instead of a few large masts.
He was careful, or at least tentative, about timing. The devices are "going into production, I think, next year", he said. No price, technical specification, supported device list or commercial rollout schedule has been announced, and Musk has a long history of missing his own production deadlines.
The satellite side of the plan
The rooftop box is one half of a hybrid network SpaceX calls Starlink Mobile, announced in October 2026. The other half is its next generation of Direct-to-Cell satellites, which let ordinary phones connect directly to satellites without extra hardware. Musk said the V2 constellation will begin launching next year with 40 times the bandwidth of the current generation.
That figure is worth noting because it differs from what he said in October, when he put the improvement at more than 100 times. The company has not explained the gap, and it may simply be Musk speaking loosely, but carriers and investors will hold him to whichever number proves right.
Building the network has already cost SpaceX heavily. It signed a 17 billion dollar deal with EchoStar in 2025 for spectrum, followed by an agreement in 2026 to acquire another low-frequency portfolio worth 8 billion dollars. According to SpaceX data, the existing Direct-to-Cell service was used by 7.4 million unique devices a month across about 30 countries as of the end of March 2026.
Whether rooftop micro cells can turn a satellite service into a genuine rival to established mobile networks is still an open question. But if the design Musk described reaches production in 2027, the dead zone in a basement flat or a steel-framed office may finally have a fix that does not depend on a signal falling from the sky.