Nobody Read the Warning Buried in SpaceX's $2 Trillion IPO: What the Market Is Missing

One dead satellite. That is all it takes.

We live in a world where a rogue satellite clipping another at orbital speed is a real scenario, not sci-fi. Closing speeds run 7 to 8 km/s. That is New York to Miami in about four minutes. At that velocity, a fragment the size of your thumbnail hits with roughly the energy of a hand grenade. The first collision throws off hundreds of new fragments. Each one becomes another grenade crossing the same crowded shell where most of the world's satellite internet already flies, about 550 km up.

Satellites start dodging. Dodging burns fuel. Fragments too small to track never show up on anybody's screen until something else breaks apart. Tens of fragments become hundreds. Hundreds become thousands. The shell starts eating itself.

SpaceX knows this. When the company went public in June 2026 in the largest IPO in history, it filed an S-1 with the SEC. That document exists so no investor can later say they were never told. Deep in that filing, SpaceX's own lawyers warn that a collision or debris-generating event "could result in a cascading series of collisions that could render certain orbital altitudes unusable" and that this would "materially and adversely affect the business."

The lawyers already settled whether a runaway cascade is possible. The question that matters now is the odds. And those odds are getting worse, fast.

Why a $2 trillion company is writing this down

Starlink already connects more than 10 million customers on ranches, islands, ships, and war zones. Places that never had a real connection. SpaceX and Blue Origin are both racing to put data centers in orbit. That buildout assumes orbits stay usable.

Since June 2026, SpaceX has been a roughly $2 trillion public company with plans to launch up to 1 million satellites. In 2019, the whole planet had roughly 2,000 working satellites. Seven years later, SpaceX alone flies more than five times that number. Nobody has built anything in space that fast.

Right now there are about 16,000 active satellites. Roughly 10,000 are Starlink. One company owns about two-thirds of every working machine over your head. SpaceX has proposals for space-based computing constellations that scale to a million satellites: data centers in orbit, where the sun hits solar panels around the clock and you never fight anyone for grid power. A million satellites is roughly 60 times everything currently flying, proposed by one company.

Blue Origin is chasing the same prize with Project Sunrise and 52,000 satellites. Rocket Lab is racing into the same economy. This is a gold rush into the same thin slice of sky.

Everyone wants the same floor

Orbits work like floors of a building. Park at a certain altitude and you stay on that floor, circling again and again. Everyone is fighting over the low floors between roughly 500 and 1,000 km. Low means a shorter signal round trip and faster internet. Low is also cheaper to reach with a rocket. So the most valuable hardware humanity has ever launched is piling into a few hundred kilometers of altitude. The sky we have in 2026 is a rounding error next to the sky these companies have already asked permission to build.

Somebody already did the math in 1978. NASA scientists Don Kessler and Burton Cour-Palais published a paper that year. Kessler ran NASA's orbital debris program. The mechanism is brutally simple. Objects hit. They shatter into hundreds or thousands of fragments. Every fragment becomes a new object moving just as fast. More objects mean more collisions. More collisions mean more fragments.

The only cleanup crew is the atmosphere. Even hundreds of kilometers up, a faint wisp of air drags debris down over years until it burns on re-entry. You are in a race: is the sky making debris faster than the atmosphere can flush it, or slower?

Kessler's warning was that there is a density threshold. Below it, orbit cleans itself and mistakes get forgiven. Above it, collisions feed themselves. You could stop launching forever and the debris count would still grow for decades, maybe centuries. That is the Kessler effect. It is a 1978 NASA paper, decades before the movie Gravity turned it into a plot. The paper is scarier.

The crash clock is collapsing

We can track about 40,000 objects, roughly softball-sized and up. Models put the real count above 10 cm closer to 54,000. Between 1 and 10 cm: about 1.2 million. Millimeter-class paint flecks and solar panel slivers: around 140 million. At 8 km/s, every one of those is a bullet. Everything under about 10 cm is effectively invisible. Nobody can dodge what nobody can see.

The worst number is the crash clock. Researchers ask: if every satellite stopped dodging, how long until the first collision between two tracked objects? In 2018 the answer was 164 days, about five and a half months of margin. By June 2025 it was down to 5 days. By May 2026 it was 2.5 days. At that level, the math says roughly a 30% chance the first collision happens inside 24 hours if the dodging stops.

Why hasn't it happened? Because the dodging never stops. Starlink satellites performed about 300,000 collision avoidance maneuvers in 2025 alone, up roughly 50% in a single year. Somewhere over your head right now, a satellite is firing thrusters so it does not hit something else. On average that happens every couple of minutes, around the clock, fully automated. The margin of safety at these altitudes is already gone. What holds the system together is software: tens of thousands of machines moving out of each other's way with no pause button.

What stops the dodging

Accidents. A healthy satellite goes quiet, then breaks apart. Tens of tracked pieces replace one controllable machine. Debris does not have software. The safety system assumes both objects in a close pass can move. The moment one cannot, the assumption fails.

The sun. In February 2022, SpaceX launched 49 Starlink satellites. A geomagnetic storm heated the upper atmosphere as they deployed. Drag spiked. Roughly 38 burned up within days. In May 2024, the strongest solar storm in about two decades degraded orbit predictions for days while thousands of satellites maneuvered at once. You cannot dodge what ground systems cannot see coming.

Ordinary failure adds more. Batteries die. Radios fail. Flight computers quit. A small percentage of tens of thousands of machines is still a lot of hardware nobody can control.

Then the darker cases. A satellite is a computer with a thruster. Take control of one and you own a machine that can burn through the densest shells humans have ever built. Kinetic anti-satellite tests have already proven how ugly this gets. China's 2007 test against Fengyun-1C at roughly 860 km created over 3,000 trackable fragments; most are still circling almost two decades later. Russia's 2021 shot at Cosmos 1408 produced roughly 1,500 trackable fragments and scared the ISS crew into their return capsules. The United States banned kinetic ASAT tests in 2022. That ban binds the country that announced it. China never signed. Neither did Russia.

Three plans, none of them enough yet

Plan one is rules. In 2022 the FCC tightened disposal: a dead satellite must come down within five years of ending its mission, down from a 25-year guideline. ESA wants net-zero space debris by 2030. That does nothing for junk already up there.

Plan two is software. Starlink dodges autonomously and conservatively. SpaceX moves when collision odds get near a few in 10 million. Old habit was closer to 1 in 10,000. Blind spot: dead satellites, hacked satellites, and shrapnel. Software cannot steer any of them.

Plan three is cleanup. ESA's ClearSpace-1 aims to capture one piece of leftover junk and drag it down. Japan's Astroscale already inspected a dead rocket stage up close. These missions remove roughly one object each, often at nine-figure cost, while industry launches thousands of new satellites a year. One down, thousands up. At that pace you never catch up.

SpaceX's deliberate parking spot helps. At roughly 550 km there is still a whisper of atmosphere. Drag bleeds off speed. Orbit decays. Hardware falls and burns in about five to six years. It is a self-cleaning orbit. That physics is real and it works in our favor.

Who loses if it goes wrong

SpaceX, by a mile. Starlink is the cash cow bankrolling Starship. Blue Origin wants Project Sunrise in that same shell. SpaceX's million-plus satellite plan for orbital compute sits on the same bet: that the floors stay usable.

Skeptics have a case. NASA's LEGEND debris model, run forward under current mitigation, projects roughly linear debris growth through 2100, not a sudden runaway. Most operators are getting dead hardware out on schedule. Serious researchers will tell you a lot of things have to fail at once before you get a true cascade. If they are right, the near-term runaway odds are very low.

But the volume is about to explode. Starship aims to make rocketry fully reusable, flying the same booster multiple times a day. Cheaper access means more companies putting hardware up. Air travel is extremely safe and planes still crash, because the fleet is huge. Scale changes the math.

Humanity has taken dangerous leaps before and come out stronger on the other side. That is the bet. The warning buried in SpaceX's IPO filing is not that the cascade is certain. It is that the company building the densest constellation in history already told investors what happens if the dodging ever fails.

Check the video here.