Starship 14: the perimeter cinema always leaves out

By the time you read this, Starship Flight 14 has either completed the first orbit in the programme’s history or it has not. Either way, the part of it that interests me happened four days before the countdown, when somebody had to sign an order closing a road.

Thirteen flights and not one orbit

Worth putting the facts on the table before the opinions. Flight 14 was set for 28 September 2026, with a seventy-five-minute window opening at 7:15 in the morning Texas time — 14:15 in Spain — and it depended on regulatory approval right up to the end.

The previous thirteen flights were suborbital: up, halfway across the planet, down. This was the first one meant to actually insert into orbit. The published plan called for deploying 26 Starlink V3 satellites, a maximum altitude in the region of 275 kilometres, about six complete orbits and close to ten hours of flight, with a single Raptor relighting in space to force re-entry. Neither the booster nor the ship was to be caught by the tower: Super Heavy was to splash down offshore in the Gulf of Mexico, and the ship in the Pacific, west of Chile.

One detail says more than all of the above. Three of those satellites carried cameras to photograph the heat shield during re-entry, and two of the shield tiles were pieces recovered from the previous flight and fitted again. That is not a launch: it is an instrumented test. The whole flight was designed to bring data home, not to get anywhere.

A launch is, before anything else, a perimeter

This is where my job and this news story meet.

For that rocket to leave the ground, somebody has to close State Highway 4 and Boca Chica Beach, which is the only way in and out of the area: roughly six miles from the checkpoint to the sand. Closure orders are usually published two to five days ahead, and since 1 September 2025 they are signed by the City of Starbase, not Cameron County, because the town incorporated in May that year and took over the job. They are published with a primary date and backup dates, because a launch slips off the calendar with an ease most people do not imagine, and the perimeter has to move with it.

It sounds like paperwork. It is not. Setting a perimeter is the hardest technical decision in an operation and it is almost never understood from outside. Intuition says you cordon off where the danger is. Practice says the opposite: you cordon off where the danger could end up, which is a much larger area, far more inconvenient, and far worse received by the people living inside it. A correct perimeter always looks excessive on the day nothing happens. And on the day something happens, it always looks too small. Those are the only two criticisms you will get, and you will get both.

The second thing nobody mentions is that a perimeter is not held up by barriers, it is held up by people. There is always someone who has to get through: the person who lives inside, the person who works inside, the person who left something inside, and the person who simply wants to see the rocket closer than everybody else. That negotiation, repeated a hundred times along six miles of road, is the actual work. The countdown is the easy part.

And if you are interested in how what happens inside a perimeter actually gets written down, I have a case file that never reached a courtroom and only circulates among my subscribers. You can ask for it here.

In the movies, nobody closes the road

Now put that next to any space launch you have ever seen on a screen.

The rockets of fiction lift off from empty deserts, from ocean platforms without a single vessel in sight, from bases where nobody seems to live. There are no notices to airmen, no maritime exclusion zone, no neighbour arguing with an officer at six in the morning because he has to get to work. The risk exists in the script, but it is treated as set dressing: an alarm, a red light, a countdown and a room full of young people in headsets.

I do not say this as a purist complaint. I say it because that omission has a very concrete narrative cost: it erases every person who makes the event possible. A real launch mobilises hundreds of people whose job consists, literally, of making sure nothing happens. Fiction keeps the ten who applaud in the control room and deletes the ones on the road, at sea and on the radio net. And then we are surprised that the future they show us looks like a set, because it is: a future with nobody holding the perimeter is a future without friction, and without friction there is no plausibility.

The problem is not the launch, it is the re-entry

There is an asymmetry in this flight worth noticing. The launch happens in a specific place, with an identifiable person responsible, a road with a name and a signed order. The re-entry happens in the middle of the Pacific, west of Chile.

When a vehicle this size enters the atmosphere and breaks up over the ocean, controlled or not, the perimeter stops being a road and becomes a strip of sea hundreds of kilometres across that is announced but not closed. Nobody posts an officer out there. A warning goes out to shipping and you trust it reaches whoever needs it. That is the exact difference between a space where someone is in charge and a space where there are only coordinates.

And that is precisely the gap that interests me as an author: the places where the infrastructure already exists and the authority has not arrived yet. I wrote about it when I looked at data centres in orbit, and it is the ground Miguel Reyes moves through in Zero Frequency: a planetary network that works flawlessly precisely because nobody has jurisdiction over it.

What was filmed and what was not

Whatever the cameras caught on Monday — six orbits, twenty-six satellites releasing one after another and two splashdowns, or an early end to all of it — it was spectacular and it was well filmed. That part never fails.

What nobody filmed is the closure order signed four days earlier. Nor the officer arguing at the Highway 4 checkpoint with somebody who had a perfectly legitimate reason to pass. Nor the strip of Pacific where nobody could guarantee anything. And it is those three things, not the rocket, that decide whether this can be done again the following week.

So here is the question I am left with: how much critical infrastructure are we willing to put in places where, if something goes wrong, there is absolutely nobody to call?

© 2026. All rights reserved