In 1978, NASA astrophysicist Donald Kessler published a paper arguing that once the density of objects in low Earth orbit passed a certain threshold, collisions between them would generate more debris than atmospheric drag could remove — and that debris would trigger more collisions, and those collisions more debris, until entire bands of orbit became too hazardous to fly through. The model did not require a rocket war or an anti-satellite weapon. It required only arithmetic. And by every public measurement available in July 2026, the arithmetic has already tipped.
The European Space Agency now estimates more than 140 million objects smaller than 10 centimetres are circling the planet, most of them too small to track, all of them moving at roughly 7 kilometres per second — about ten times the muzzle velocity of a rifle round.
What Kessler actually predicted
Kessler’s 1978 paper did not claim a chain reaction would happen next Tuesday. It claimed something subtler and worse. Above a certain density, the rate of new debris created by random collisions would exceed the rate at which atmospheric drag pulled old debris down. The population would grow on its own, without a single new launch.
The threshold was a tipping point, not a cliff. Once crossed, the cascade would play out over decades. Satellites would fail. Replacements would launch into a dirtier sky. Each generation of hardware would face a slightly higher collision probability than the one before it. The bands most affected — roughly 700 to 1,000 kilometres up, where sun-synchronous imaging satellites live — would eventually become unusable for any spacecraft not built like a tank.
The 1978 paper assumed a satellite population in the low hundreds. As of mid-2026, active satellites number in the tens of thousands, and a single operator flies close to 10,000 of them.
Why a paint fleck matters at orbital velocity
Kinetic energy scales with the square of velocity. A fragment 5 centimetres across, moving at 7 kilometres per second, carries roughly the energy of a small car hitting a wall at highway speed. A fleck of paint can crack a shuttle window. A bolt can end a mission.
Eric Felt, a retired U.S. Space Force colonel now at the University of Texas, has explained that small objects traveling at orbital velocities can cause catastrophic damage to satellites. His team is working with the Florida startup SOAR — whose advisor is Donald Kessler himself — on a passive shielding satellite designed to absorb sub-10-centimetre debris that ground radars cannot see. That partnership was announced on 15 July 2026.
The tracked catalogue — objects large enough for the U.S. Space Surveillance Network to see and follow — contains something like 40,000 items. The untracked population is roughly 3,500 times larger.
The threshold, and the year it was crossed
Different research groups place the tipping point in slightly different places, but all of them are now behind us. Analysts have argued since the mid-2000s that the 900-to-1,000-kilometre altitude band was already unstable — meaning collisions there would generate more debris than drag removed even if launches stopped tomorrow.
Two events sharpened the picture. In 2007, China destroyed its own Fengyun-1C weather satellite with a ground-launched missile, creating more than 3,000 trackable fragments in a single afternoon. In 2009, the defunct Russian Kosmos 2251 satellite collided with the active Iridium 33, generating roughly 2,000 more. Those two events alone increased the trackable debris population in low Earth orbit by about a third.
Every fragment is now its own potential bullet, orbiting for decades before drag brings it down. And the fragments from a collision at 800 kilometres will circle the planet for a century or more.
What megaconstellations changed
When Kessler wrote his paper, the total number of operational satellites was fewer than 500. By July 2026, one commercial network alone accounts for roughly 10,000 spacecraft, and filings with the International Telecommunication Union propose hundreds of thousands more from a handful of operators. Ars Technica reported in July 2026 that low Earth orbit is now dense enough that companies are reviving the idea of LEO navigation constellations to supplement GPS, precisely because the sky up there is now crowded enough to support such a service.
Defenders of the megaconstellation model point out that most of those satellites fly below 600 kilometres, where atmospheric drag will de-orbit a dead spacecraft within about five years without intervention. That is true. It is also the reason 550 kilometres has become the most valuable altitude band in the solar system — and the reason it is filling up fastest.
The problem is not the operators who fly low and de-orbit responsibly. The problem is what happens when a satellite at 550 kilometres suffers a fragmentation event and throws pieces into higher, longer-lived orbits. Or when a rocket body left in a 700-kilometre parking orbit breaks up decades after its launch. LeoLabs and independent debris analysts have flagged China’s practice of abandoning upper stages in orbit as one contributor; unidentified metallic spheres that washed up on Forrest Beach in Queensland in July 2026 were traced by the Australian Space Agency to a foreign rocket body that had recently re-entered.
The debris you cannot see
The most dangerous fragments are the ones nobody is tracking. Ground radars can reliably see debris down to about 10 centimetres in low Earth orbit; at geostationary altitude, roughly 36,000 kilometres up, the practical limit is closer to a metre. Everything smaller is invisible until it hits something.
A study used reprocessed telescope archives to find previously undetected fragments in geosynchronous orbit as small as 5 centimetres. Some had apparently been circulating for years. The researchers noted that these objects — hidden in the background noise of older images — could destroy an active satellite on impact and would remain in that orbit essentially forever, because there is no atmospheric drag to speak of at 36,000 kilometres.
ESA’s 140-million figure covers only fragments larger than a millimetre. Below that, the population is functionally uncountable.
Why the sky is being handed out first-come, first-served
The 1967 Outer Space Treaty forbids any nation from claiming sovereignty over orbit. It does not forbid a company from filling that orbit with hardware. The International Telecommunication Union, a UN agency in Geneva, coordinates radio-frequency and orbital-slot allocations on a first-to-file basis — a system designed in an era when a handful of governments launched a handful of satellites per year.
That system is now being used to book hundreds of thousands of slots by a small number of private operators. The U.S. Federal Communications Commission, which authorises American commercial launches, recently approved Reflect Orbital’s Earendil-1 mirror satellite despite receiving more than 1,600 public comments raising concerns about astronomy, wildlife and orbital sustainability. The agency’s response, as summarised by University of Texas professor Moriba Jah, was that those concerns fell outside its jurisdiction over radio spectrum.
Jah told Newsweek the pattern was familiar: regulators often approve projects without fully assessing long-term orbital debris consequences. Currently, U.S. federal agencies do not review the environmental or orbital-debris implications of new commercial constellations as a matter of statutory requirement.
What a cascade actually looks like
Hollywood versions of Kessler Syndrome — the 2013 film Gravity being the most familiar — compress the cascade into minutes. Real cascades unfold across years and decades. A collision creates a debris cloud. That cloud spreads along the orbital plane. Over months, orbital perturbations smear it around the Earth. Over years, individual fragments encounter other satellites. Each encounter has a probability, not a certainty, of producing another fragmentation.
The endpoint is not an empty sky suddenly filled with a Bruckheimer explosion. The endpoint is insurance premiums that make certain orbits uneconomic. Then it is operators avoiding those altitudes entirely. Then it is scientific missions — Earth-observation, climate monitoring, radio astronomy — quietly cancelled because the risk-adjusted cost of flying them exceeds their budget. The sky does not close. It just becomes a place where only the richest and most expendable hardware goes.
The cleanup problem
Removing debris is technically possible and economically brutal. Proposed methods include harpoons, nets, magnetic capture, robotic arms and — in a NASA-studied concept — ground-based lasers that nudge fragments into lower orbits by ablating a tiny amount of material off one side. All of these work in principle on debris large enough to see. None of them scale to the 140 million uncatalogued fragments smaller than 10 centimetres.
The SOAR–UTEP passive-shielding approach takes the opposite tack: instead of chasing individual fragments, fly a Whipple-shielded sacrificial structure through a heavily contaminated orbit and let the debris hit it. Sensors record each impact. The satellite absorbs momentum. Some fragments are captured; others are broken into pieces small enough to de-orbit quickly on their own.
Nobody involved thinks this cleans up low Earth orbit. As Felt noted, cleaning up all debris in low Earth orbit is impractical; efforts must focus on protecting the most critical orbital regions. Triage, in other words.
What is already at risk
Low Earth orbit hosts the International Space Station, the Chinese Tiangong station, the entire Earth-observation industry, most weather satellites, and the growing communications constellations that provide internet to remote regions. GPS sits higher, at about 20,200 kilometres, but its ground-augmentation and timing systems depend on LEO relay hardware.
The station Skylab, whose 1973 emergency rescue saved America’s first orbital laboratory, flew at roughly 435 kilometres — an altitude that today sees regular debris-avoidance manoeuvres by the ISS. In November 2021, a Russian anti-satellite test forced the station’s crew to shelter in their return capsules while a debris cloud passed. That was one test, one satellite, one afternoon. The cloud is still up there.
Newer entrants keep arriving. In July 2026, a Miami startup called City Labs launched the first commercially built nuclear-powered CubeSat as a rideshare payload — a softball-sized spacecraft joining a queue of small satellites that must now be tracked, catalogued and, eventually, disposed of.
The century-long clock
A satellite at 400 kilometres re-enters within a few years. A satellite at 600 kilometres takes about 25 years. A satellite at 800 kilometres takes roughly 100 years. A satellite at 1,000 kilometres takes more than 1,000 years. There is no wind at those altitudes to speak of. There is only the faintest whisper of residual atmosphere, and it is not enough.
Every fragment created above 700 kilometres in 2026 will still be there when the last human alive today is long dead. The Kosmos-Iridium fragments from 2009 will outlive every reader of this article, every reader’s grandchildren, and the institutions that currently regulate them. Kessler’s 1978 model assumed a debris population that would grow slowly. The population grew fast. The threshold was crossed quietly, without a headline, sometime in the last two decades.
The satellites moving overhead tonight — the ones that look like stars but drift — are the visible tenth of a percent. The rest is invisible, moving at seven kilometres per second, and waiting.
