The Band That Goes Through Walls: SpaceX Buys Its Way Into Cellular

8 hours ago

6 min read

A smartphone on a concrete ledge in a dark parking garage, sending warm amber radio waves straight through the concrete pillars, with satellites glinting through a ceiling vent above.

The FCC just opened its review of SpaceX's $8 billion purchase of 800 MHz spectrum. The airwaves matter. The price matters more, because it tells you what the market thinks a satellite company is worth as a phone company.

What the FCC actually did

On November 10, the FCC's Wireless Telecommunications Bureau issued a public notice accepting for filing SpaceX's applications to take control of two Grain Management license companies, Prism License Co. 1 and Prism License Co. 2. The buyer is Starlink Spectrum Holdings, a wholly owned SpaceX subsidiary. The docket is WT 26-313.

Acceptance is not approval. It means the paperwork cleared an initial review and the public gets to argue. Comments and petitions to deny are due December 1. Oppositions are due December 8.

That distinction is about to get flattened in a thousand headlines, so hold onto it. Nothing has been granted. What changed is that the objection window is open, and Verizon, AT&T, and T-Mobile can now formally explain why this shouldn't happen.

Why 800 MHz is the whole ballgame

SpaceX already owns spectrum. The FCC approved its purchase of EchoStar's airwaves on May 12, a deal worth roughly $40 billion and split with AT&T. SpaceX's share is about 65 MHz, sitting up around 2 GHz.

Two gigahertz travels fine. It doesn't go through walls especially well, and it really doesn't like being indoors. 800 MHz does. That's the band that made cellular work in the first place, the one that reaches basements, parking garages, elevators, and the inside of a house in a way higher frequencies struggle with. It's why 14 MHz matters more than it sounds.

The portfolio covers 546 paired licenses across 3,224 counties, roughly 100% of the U.S. population, across all 50 states, D.C., Puerto Rico, and the Virgin Islands. The frequencies sit at 817 to 824 MHz and 862 to 869 MHz, and most phones already sold support them. Nobody has to buy a new handset.

Musk called it "the last critical piece of the spectrum puzzle needed for SpaceX to provide complete phone coverage in America."

That's marketing. It's also basically correct about the physics.

The same airwaves, twice the price

The licenses came from T-Mobile, which had to divest its 800 MHz holdings as a condition of the 2020 Sprint merger. T-Mobile sold the portfolio to Grain Management, a transaction completed August 11, for $2.9 billion in cash.

SpaceX agreed to pay roughly $8 billion for it on October 8.

Same licenses. Same counties. Roughly two months apart, and a markup on the order of 2.75 times, for a private equity firm that did nothing but hold the asset and file paperwork.

That's the number I keep coming back to, because it isn't really a spectrum story. T-Mobile valued those airwaves as a marginal asset it had been ordered to sell. SpaceX valued them as the difference between a satellite service and a phone company. Both can be right.

It also tells you something about the bidding. In August, SpaceX and AST SpaceMobile were both reported as circling Grain's portfolio when it was valued nearer $6 billion. By October, one of them had agreed to pay retail plus.

The spectrum math doesn't make SpaceX a monopoly

Here's the part that should calm everyone down, at least on antitrust grounds.

After the deal closes, SpaceX would hold up to 89 MHz of spectrum in any given market, including up to 14 MHz below 1 GHz. The FCC's spectrum screen, the threshold that flags a transaction for closer competitive review, sits at 385 MHz. SpaceX would land at less than a quarter of it.

AT&T, Verizon, and T-Mobile each hold roughly 300 MHz.

So this isn't a company buying market power. It's a company buying coverage, which is the one thing it can't manufacture from orbit. The licenses are thin, low, and everywhere, which is precisely what you want if your business model is that your phone keeps working when there's no tower nearby.

Spectrum is now the easy part

Which brings up the part the headlines will skip.

SpaceX has spectrum. It doesn't have a network. In early November it picked Qualcomm to supply the radios, baseband units, and software for a nationwide 5G build, with deployment across tens of thousands of cell sites starting in 2027. One trade publication put the target at 80,000 sites. Starlink president Gwynne Shotwell has said service on SpaceX's own airwaves is targeted for early 2028.

Tens of thousands of sites means land, leases, zoning, power, backhaul, permits, and fiber, in thousands of jurisdictions, most of which have never dealt with Starlink and some of which will be delighted to make the process interesting. The incumbents spent decades and hundreds of billions of dollars on exactly that, and their towers are half the reason their spectrum is worth anything.

That's why the tower companies were the quiet winners when this deal was announced. Somebody has to own the steel.

There's also a separate application SpaceX filed on October 27 seeking authority to provide supplemental coverage from space on these same frequencies, plus waivers. That's still pending. The November 10 notice doesn't decide how SpaceX can use the spectrum from orbit. Ground licenses and space authority are two different permissions, and SpaceX needs both.

And the buildout conditions attached to Grain's July purchase from T-Mobile, which require rapid construction for both terrestrial and direct-to-device service, carry over to SpaceX if the transfer is approved. Conditions like that have teeth. Ask anyone who has ever held a license with a coverage deadline.

What it means for the Big Three

Three things, in order of how much they should worry.

First, pricing power at the margin. Verizon, AT&T, and T-Mobile make a lot of money where coverage is hard, which is rural America, building interiors, and anywhere a tower is expensive. A competitor that reaches those places from the sky, through low-band spectrum, using the phone already in the customer's pocket, attacks the highest-margin part of the business first.

Second, wholesale. If Starlink Mobile builds a real network, it can sell access to cable companies and MVNOs that currently buy from the Big Three. That's how you become a carrier without winning retail customers one at a time, and it's the scenario the incumbents are least equipped to price against.

Third, the regulatory drift. The FCC voted on October 29 to advance a proposal opening another 482 MHz to direct-to-device service. The Big Three pushed to strip nearly 300 MHz out of it. They aren't really fighting about this deal. They're fighting about the next one, and about whether satellite ends up a permanent competitor or just a roaming partner.

Musk added fuel on November 6 at the Baron Investment Conference, saying SpaceX could buy a legacy telecom carrier and use its spectrum "ten times more effectively." Nobody has named a target. The comment alone moved stocks.

The market reaction to the original announcement was severe enough that Barron's published a piece telling investors to buy the dip. Both reactions can be right. The incumbents have roughly 300 MHz apiece, well over 100 million connections each, and physical plant that took decades to assemble. SpaceX has 89 MHz, a satellite constellation, a Qualcomm contract, and a construction schedule.

What would make me wrong

The obvious one. The FCC hasn't approved anything. Petitions to deny are due December 1, and the incumbents have every incentive to file them and to slow this down procedurally. A deal can clear the spectrum screen and still lose a year.

The second. Building a terrestrial network is a miserable business, and SpaceX has never done it. Its core competency is launching things and operating satellites, which it is genuinely extraordinary at. Running 80,000 cell sites is somebody else's company.

The third. Eight billion dollars for 14 MHz is a bet that direct-to-device becomes a mass market rather than a rural and emergency feature bundled into premium plans. If it stays the latter, that's a very expensive sliver of airwaves.

What to watch

December 1, when the petitions to deny land, and December 8, when the oppositions do.

The October 27 space-coverage application, which is the permission that actually lets this spectrum serve a phone from orbit.

Whether the Qualcomm build holds a 2027 schedule, because that's the difference between a spectrum story and a carrier. And whether anyone credible prices out what it costs to cover the places the Big Three would rather not.

The airwaves were always the scarce part. SpaceX just showed what it thinks that's worth.

Sources

Rich Washburn is a technologist, strategist, and Founder & Chief AI Architect of ARIA AI Labs, working at the intersection of AI, infrastructure, communications, and capital. He also serves as Managing Partner and Chief AI Officer at Eliakim Capital.