From the newsletter: SpaceX’s potential move into retail mobile via Starlink changes the telco game, as hybrid networks, new spectrum, and trillion-dollar valuations challenge the carrier status-quo.
You can picture the big chiefs at the big US carriers spitting cornflakes across their breakfast tables – on reading the FT report this morning that SpaceX is getting into the retail game. It was always on the cards; everyone suspected so; no one thinks Musk’s masterplan is just to fill a gap. But still, alarm bells will be ringing for US telcos now – if they weren’t already. It changes the game: Starlink goes from partner to competitor. You can’t see AT&T and Verizon changing tack to offer D2D roaming with Starlink now, and you wonder whether T-Mobile might review its position.
For the traditional carrier set, the existential angst, here, apart from the Musk factor, is that parent company SpaceX has spectrum for a terrestrial network. It paid $17 billion to EschoStar last September for about 40 MHz of paired mid-band AWS-4 spectrum (the 2000-2020 MHz block for uplink traffic, and 2180–2200 MHz block for the downlink) in the US, plus a further $2.6 billion in November for adjacent mid-band assets and extensions in the same 2 GHz portfolio. That 40 MHz holding is usable for a hybrid terrestrial/satellite offering – if it had infrastructure on the ground.
The FT says SpaceX president Gwynne Shotwell told investors at an “IPO roadshow” that Starlink is considering this. Starlink has about 10 million subscribers, apparently, but these mostly take fixed satellite broadband, and not D2D services in their phones. It is a profitable business, generating revenue of $11.4bn (about 61% of SpaceX’s total takings, across three core units) and EBITDA of $7.168bn in 2025. But even Starlink’s returns – up 49.8% and 86.2% in the year, respectively – are not enough to push SpaceX into the black.
SpaceX earned $18.674bn in revenue in 2025, but lost $2.58bn. So its record-breaking IPO valuation this month ($1.77tn) was carried by its Starlink business, plus its rockets programme (serving the whole LEO industry), and driven upwards by its pitch about a hybrid satellite/terrestrial compute-and-network play – framed within a projected $28.5tn (!) TAM, including about $1.6tn for Starlink, potentially, split about-evenly across its broadband and mobile use cases. So everyone knew.
Brokerage firm Oppenheimer, reports Reuters, said in a note earlier this month that SpaceX will disrupt the $1.6 trillion US comms industry with expansion of Starlink. That 40 MHz of AWS-4 spectrum gives Starlink roughly 10-15% of the usable sub-6 GHz spectrum base of a typical US operator – which is enough to matter strategically, but nowhere near enough to replicate a full-scale terrestrial network. So the tension is not about spectrum arithmetic, but architecture. Starlink doesn’t have to out-muscle them on ground capacity when it can ride over them in space.
The satellites provide ubiquitous coverage, the AWS-4 network offers selective infill, and the whole bundle is sold via Startlink shops (or whatever). Of course, AT&T et al might shrug it off; it said last week that satellites mop up the rural 1% – while the old guard shares-out the rest. And the physics and economics have to work: satellites can’t compete on throughput or speed. But they can be the universal fallback access layer – and reshape control of connectivity at the margins, where the experience is most fragile and consequential, and where the economics are often decided.
And so this matters, clearly. AT&T and Verizon, at least, will surely lean into AST SpaceMobile, as an alternative, and a genuine D2D operator – whose model is to extend terrestrial cellular rather than to compete with it. And Europe, reimagining its tech sector in the name of sovereignty, will be even more wary of anything that bypasses its terrestrial licensing and spectrum economics. It will be interesting to see how SpaceX’s IPO valuation stands up to the real establishment – to the legacy industries and spooked regulators that it wants to disrupt.
