Level 1 — Absolute Beginner

SpaceX is a space company. It sends rockets to space. It also has many small satellites. They give people the internet.

Now SpaceX wants to do something new. It wants to build a phone network on the ground. A phone network needs tall towers with radios on them.

SpaceX asked a company called Qualcomm for help. Qualcomm makes radio parts for phone towers. The work will start in 2027.

The plan is very big. SpaceX wants tens of thousands of towers. Qualcomm says this is the biggest deal of this kind it has ever made.

a rocket

A machine that flies into space.

a satellite

A machine that goes around the Earth in space.

the internet

The system that connects computers and phones around the world.

a network

A system of things that are connected to each other.

a tower

A tall structure, here one that holds radio equipment.

a radio

Equipment that sends and receives signals through the air.

the ground

The surface of the Earth that we stand on.

a deal

An agreement between two companies or people.

Level 2 — Elementary

SpaceX is best known for rockets and for Starlink, the service that delivers internet from thousands of satellites in low orbit. On Wednesday the company announced something that looks almost old-fashioned by comparison. Starlink Mobile is going to build an ordinary mobile phone network on the ground across the United States, and SpaceX has chosen Qualcomm to supply the equipment.

Building a national network means installing radio units on tens of thousands of cellular sites. These are the masts and rooftop installations you see beside motorways and on top of tall buildings. Qualcomm will provide two of its Dragonwing platforms, called QRU100 and QDU100, and deployment is planned to begin in 2027 under a multiyear agreement.

One technical detail explains why the deal matters. The radios use a system known as massive MIMO with 64 transmit and 64 receive antenna paths. Instead of one antenna shouting in every direction, dozens of small antennas work together and aim narrow beams at individual phones. This lets a single mast serve far more customers at once without the signal getting worse.

Qualcomm says it is the largest 5G infrastructure agreement the company has ever signed, and neither side disclosed how much it is worth. The strategic logic is clear enough. Satellites are excellent at covering empty land, but they cannot carry the traffic of a crowded city. If Starlink Mobile wants to compete with established mobile operators, it needs towers on the ground as well as hardware in the sky.

an orbit

The curved path an object follows around a planet.

terrestrial

On the ground, rather than in space or in the air.

a cellular site

A location with radio equipment that serves mobile phones nearby.

a mast

A tall pole or tower that carries antennas.

to deploy

To install and put equipment into active use.

an antenna

The part of a radio system that sends or receives signals.

a beam

A narrow directed stream of radio energy or light.

an operator

A company that runs a telephone or network service.

Level 3 — Intermediate

For a company built on the idea that infrastructure should be moved into orbit, SpaceX's announcement on Wednesday reads like an admission. Starlink Mobile has selected Qualcomm to supply the radio hardware for a nationwide terrestrial 5G network in the United States, covering tens of thousands of cellular sites under a multiyear agreement, with installation starting in 2027. The towers will be ordinary towers, bolted to the ground.

The equipment is Qualcomm's Dragonwing line, specifically the QRU100 radio unit and the QDU100 distributed unit, configured for massive MIMO with 64 transmit and 64 receive paths. The engineering significance of that number is worth unpacking. A conventional antenna radiates broadly and wastes most of its energy on empty air. A 64-by-64 array instead synthesises dozens of steerable beams, tracking individual handsets and reusing the same spectrum simultaneously in different directions, which multiplies the capacity of a single site without acquiring any new frequencies.

Qualcomm characterises the contract as the largest 5G infrastructure agreement in its history, a claim that says as much about the company's ambitions as about SpaceX's. Qualcomm's reputation rests on the modems inside phones rather than on the equipment inside base stations, a market long dominated by Ericsson, Nokia and Huawei. A customer of this size gives Qualcomm a credible reference for every network operator it approaches next. Neither party disclosed financial terms.

The strategic reasoning behind the deal exposes the real limits of satellite connectivity. Direct-to-cell from low orbit is transformative for a hiker in a canyon or a ship in the mid-Atlantic, because it converts zero coverage into some coverage. It is close to useless in Manhattan at five in the afternoon, where the constraint is not coverage but capacity, and where a satellite beam a hundred kilometres wide would have to be shared by millions of devices. A network that intends to serve cities has to be built where the people are.

an admission

An acceptance that something uncomfortable is true.

a distributed unit

The part of a mobile base station that processes radio signals for several radios.

to synthesise

To create something by combining several parts.

steerable

Able to be aimed or pointed in a chosen direction.

spectrum

The range of radio frequencies a network is licensed to use.

capacity

The total amount of traffic a system can carry at one time.

a base station

The fixed radio equipment that connects mobile phones to a network.

a constellation

A group of satellites working together as one system.

Level 4 — Advanced

There is a quiet irony in Wednesday's announcement. SpaceX, the company whose entire proposition has been that infrastructure belongs in orbit, has contracted Qualcomm to supply radio hardware for a nationwide terrestrial 5G network, spanning tens of thousands of cellular sites under a multiyear agreement with deployment commencing in 2027. Starlink Mobile, in other words, is going to do what every incumbent operator has done for forty years: acquire sites, pour concrete, and hang radios on steel.

The hardware is Qualcomm's Dragonwing portfolio, the QRU100 radio unit paired with the QDU100 distributed unit, configured as a massive MIMO array with 64 transmit and 64 receive paths. That configuration is the pivotal engineering choice. A legacy sector antenna illuminates a wide arc and dissipates most of its radiated power into unoccupied space; a 64-element array performs digital beamforming, generating dozens of narrow, independently steered lobes that track handsets and permit spatial reuse of identical spectrum. The result is a multiplication of site capacity achieved without acquiring an additional megahertz of licensed bandwidth, which in a spectrum-constrained market is the only lever that scales.

For Qualcomm the significance is commercial rather than technical. The company's franchise has always been the modem and the system-on-chip inside the handset; the radio access network has been the territory of Ericsson, Nokia and, until geopolitics intervened, Huawei. Winning an anchor customer of this scale supplies the reference deployment that every subsequent sales conversation requires, and it is why Qualcomm has been willing to describe the contract as the largest 5G infrastructure agreement in its history while disclosing none of its economics.

Underneath the announcement sits an honest acknowledgement of what satellite connectivity can and cannot do. Direct-to-cell service from low Earth orbit is genuinely transformative at the margin of coverage, where the alternative is nothing at all: a vessel beyond the shelf, a hiker in a slot canyon, a disaster zone with its terrestrial plant destroyed. It is structurally incapable of serving dense demand, because a beam footprint measured in tens of kilometres must be shared among every device inside it, and urban traffic is a capacity problem rather than a coverage problem. Physics, in the end, requires the network to be built where the subscribers actually stand.

an incumbent

An established company already occupying a market.

beamforming

Combining antenna signals so that radio energy is concentrated in a chosen direction.

a lobe

One directional region of radiated energy from an antenna.

spatial reuse

Using the same frequency at the same time in separate directions or places.

bandwidth

The width of the frequency range available for carrying signals.