The satellite company is building towers.

We love the magic trick.

Rockets landing on barges. Satellites talking directly to the phone in your pocket. Last week, I wrote about that magic trick.

Then comes the boring part.

On November 4, Qualcomm shared a quiet piece of news. SpaceX is buying cellular gear to build a 5G network on the ground in the US.

Tens of thousands of macro cell sites, starting in 2027.

Wait. Why does a satellite company need cell towers?

Because physics is undefeated. Satellites are great for coverage, for reaching the places nobody else goes. They are terrible for dense cities. A phone's tiny signal is hard to hear from space, and capacity runs out fast when thousands of people are streaming in the same neighborhood.

So SpaceX is becoming a carrier.

They are using Qualcomm's Dragonwing infrastructure — QRU100 massive MIMO radios with 64 transmit and 64 receive paths, paired with QDU100 baseband units and Layer 1 software. Normally, engineers deploy massive MIMO at 3.5 GHz and above. SpaceX plans to run it around 2 GHz, using the roughly 65 MHz of mid-band spectrum they bought across roughly $19.6 billion in EchoStar deals. (A Qualcomm executive called doing this at 2 GHz an innovation.)

Add the low-band 800 MHz spectrum SpaceX bought in October. Add the FCC voting in late October to open 482 MHz more spectrum for satellite-to-phone service, plus proposing a 25 MHz auction for 5G or direct-to-device.

You have seen the $100 billion headlines. To be plain: that is not a confirmed budget from SpaceX. It is an analyst estimate from Joe Madden at Mobile Experts (via Axios) for the combined satellite and ground network, including spectrum. Actual financial terms with Qualcomm were not disclosed.

And right now, nothing is live. No coverage map. No price plan. No subscribers.

Still, look at the turn.

For years, the industry argued about whether satellites would replace cell towers. SpaceX's answer was to build both.

It is easy to fall in love with the exciting part of your work. The rockets. The big launch.

The real advantage belongs to the people willing to do the boring part, too.