When the astronomer Samantha Lawler moved to a farm in rural Saskatchewan seven years ago, the darkness overhead was part of the draw. From her homestead she could see the kind of sky most city dwellers only imagine, crowded with stars. Then the stars began to move. “Now there are many satellites all the time,” Lawler told Scientific American, whose feature on the crowding of low-Earth orbit anchors this piece. “I really could notice the change that was happening.”
That change has a name and a date. In May 2019, the same year Lawler relocated, SpaceX launched its first 60 Starlink satellites. This past March the company sailed past 10,000 active satellites and 10 million paying customers. Starlink alone now accounts for roughly two-thirds of every working spacecraft in orbit. And reading across the GeoMemo Intelligence Dataset, the age of the megaconstellation is only getting started.
The ramp
The growth is the shape engineers call a hockey stick. A trickle of experimental satellites in 2019 became a steady drumbeat, then a flood. By the start of 2026 the sky held something close to 15,000 working satellites, and Starlink was most of them. What took the entire space age to assemble, a few thousand functioning spacecraft, SpaceX rebuilt and doubled in a handful of years.
The land-grab
Starlink is not racing alone, and the paperwork tells the story better than any launch does. In January 2026 the U.S. Federal Communications Commission cleared SpaceX to expand toward 15,000 satellites; by July the company had filed for a third-generation fleet of 100,000, and separately floated the idea of a million orbiting artificial-intelligence data centers. China answered in kind, its government filing for roughly 200,000 satellites across the Guowang and Qianfan programs. Amazon, Blue Origin, Europe and a scattering of startups have queued up behind them.
Add every application together and the number of satellites now sitting in regulators’ inboxes approaches two million. Not all of those will fly; some filings are best understood as first-mover land claims, a way to reserve orbital shelf space before rivals can. But the direction is unmistakable, and it worries the people who write the rules. “A million satellites? Give me a break,” the satellite-regulation expert Ruth Pritchard-Kelly told Scientific American, calling SpaceX’s filing a possible exercise in shock and awe.
The stakes are not only commercial. In Foreign Policy, the analysts Robert Muggah and Misha Glenny argued that Starlink has effectively “privatized geopolitics,” turning a single company into “a gatekeeper in orbit, helping decide who connects as well as where, under what conditions.” That gatekeeper has already shaped a war in Ukraine and unrest in Iran, and it has drawn the attention of rivals. Chinese researchers have modeled electronic-warfare attacks on Starlink; Russia is reported to be developing weapons to blind it.
The next wave
The giants are not the only ones coming. Behind Starlink and China’s programs, a second tier is taking shape, and it divides into three kinds of players.
The first are niche operators already flying hardware. AST SpaceMobile, with its BlueBird satellites, and Lynk Global are building direct-to-cell networks that connect an ordinary phone straight to a spacecraft, no dish required. Their fleets are small, a handful of satellites apiece, but the hardware is real and in orbit; between them they surface in more than 300 reports across the GeoMemo Intelligence Dataset.
The second are national champions, still mostly on paper. Russia is the clearest case. The private venture Bureau 1440 is assembling project Rassvet, roughly 292 satellites in its first phase, with a stated goal of some 900 by the mid-2030s; the state Sfera program sketches another 600 across broadband and internet-of-things layers. Six of Rassvet’s satellites are actually up; barely a handful of Sfera’s are. Rivada Space Networks, for its part, holds a license for 600 and has launched not one.
The third, and the reason the first two matter, are the enablers. India does not yet run a Starlink rival, but ISRO is a serious launch power, and two Indian startups, Skyroot with its Vikram rocket and Agnikul with Agnibaan, are driving the cost of reaching orbit steadily down. Rocket Lab is doing the same in the West. Cheap launch is what turns a filing into a fleet.
So, are these megaconstellations in the making? Not one of them is a megaconstellation today, and some never will be. But two are already launching, and the barrier to becoming the next one is falling fast. The sky is not filling from the top alone, and every new entrant, from Starlink down to a six-satellite startup, is aiming at the same finite shells.
The junkyard
Everything launched must eventually come down, or stay up forever. Right now, far too much is choosing forever. There are roughly 36,000 tracked pieces of debris larger than ten centimeters circling Earth, and tens of millions of smaller fragments, together weighing something like 13,500 tonnes. They travel at seven to ten kilometers a second, fast enough that a fleck the size of a paint chip can punch through a working satellite.
The nightmare scenario has a name: Kessler syndrome, a runaway cascade in which one collision spawns debris that causes the next, and the next. It has happened once already, when the Iridium and Cosmos satellites smashed together in 2009 and left more than 2,000 trackable pieces, roughly half still aloft. Experts now believe the tipping point is behind us in the busiest lanes. “We have already exceeded the critical number of objects that can be safely put into low-Earth orbit,” the University of Birmingham debris specialist Hugh Lewis told Scientific American; above 550 kilometers, he says, the debris population keeps growing even if every launch stopped tomorrow. Higher still, at 2,000 kilometers, junk can linger for 100,000 years, a span comparable to the entire history of our species.
“We are ripe for a major event to occur. And all of that debris will rain down through all the other operational satellites.”Darren McKnight · LeoLabs, in Scientific American
The chaos math
So how full can orbit actually get? No one can say, and that is the problem. A 2022 study put the theoretical ceiling at 12.6 million spacecraft; a 2024 follow-up widened it to somewhere between 10 and 100 million, a range so broad it offers policymakers nothing to hold on to. The satellites themselves are already doing frantic work to stay apart. Starlink performs on the order of a thousand collision-avoidance maneuvers a day, and told regulators it had dodged some 300,000 possible collisions in a single year, about forty per satellite.
Push the numbers to Elon Musk’s vision and they stop making sense. Lewis calculated that a million-satellite constellation would demand roughly a billion avoidance maneuvers a year. “It’s ridiculous,” he said. “You can’t grasp that. There aren’t that many seconds in a year.” And the ceiling is not fixed: research by his Birmingham colleague Matthew Brown suggests that greenhouse gases, by cooling and thinning the upper atmosphere, could cut the natural drag that drags dead satellites back down, roughly halving how much orbit can safely hold. We may be shrinking the parking lot even as we fill it.
The costs you cannot see
The crunch is not only about collisions. For astronomers it is an unfolding catastrophe of light. The Atacama Desert’s Paranal Observatory, future home of the most powerful optical telescope ever built, is fighting off light pollution and nearby industry, while the European Southern Observatory’s images now show satellite streaks slashing across the frame where meteors used to fall. When Australia cut its astronomers’ access to those telescopes, the Nobel laureate Brian Schmidt called it a “hammer blow.” And the FCC has now approved Reflect Orbital’s plan to fly mirrors that beam sunlight to Earth after dark, over the objection of researchers such as Lawler and Aaron Boley, who warn in The Independent that the reflections could “overload and fry” telescope detectors.
There is an atmospheric bill, too. One study cited by The Independent’s Andrew Griffin projects that rocket launches could account for 42 percent of the space sector’s climate impact by the decade’s end, with black carbon high in the stratosphere hundreds of times more potent than soot at ground level. What happens as thousands of satellites burn up on reentry each year, seeding the sky with vaporized aluminum, nobody yet knows. And the losses are cultural as well as scientific: as the stars fade, so does an inheritance. “We are the only culture in the world that looks for dark spots in the Milky Way,” the Yorta Yorta man Kai Lane told SBS News, describing sky-knowledge that light pollution is quietly erasing.
The rules we never wrote
Here is the twist: the limit on all of this may prove to be legal, not physical. The governing framework is still the 1967 Outer Space Treaty, drafted in an era of a few dozen government satellites and utterly silent on private megaconstellations. There is no orbital traffic controller, no binding right-of-way. Coordination between operators was, until recently, conducted by email and phone; at the scale now arriving, that is simply unsustainable. Voices from the International Telecommunication Union to Foreign Policy are calling for something like a “Space COP,” a global forum to govern orbit before it seizes up. “They need rules before orbit gets crowded,” the ITU’s Alexandre Vallet warned.
None of this makes space unusable, and it is worth being honest about the upside. The same constellations deliver broadband to villages, ships and disaster zones that had none, and let soldiers fly drones from half a world away. “It’s not that space will be unusable,” the Aerospace Corporation’s Brian Weeden told Scientific American. “It’s just a question of what costs people are willing to bear. There might be some orbits that become too costly for pretty much anyone to operate in.” That is the real question hanging over the space jam. It was never whether we can fill the sky. It is whether we should, how much, and who gets to decide. Until someone writes those rules, the honest answer to Scientific American’s question, is chaos inevitable, is that we are choosing not to find out.
Common questions
How many satellites are in orbit right now? Around 15,000 working satellites circled Earth at the start of 2026, and Starlink alone accounts for roughly two-thirds of them.
How many satellites can low-Earth orbit safely hold? No one knows, and that is the problem. Peer-reviewed estimates range from about 12.6 million to somewhere between 10 and 100 million, a spread far too wide to guide policy.
What is Kessler syndrome? It is a runaway chain reaction in which one orbital collision creates debris that triggers the next, and the next. Many experts believe it has already begun in the most crowded lanes above 550 kilometers.
Who is building satellite megaconstellations besides Starlink? China’s Guowang and Qianfan programs are the largest rivals, and a second tier is forming behind them, including AST SpaceMobile, Lynk Global, Russia’s Rassvet and Sfera projects, and Rivada.
Who is covering this
The reporting behind this piece spans dozens of newsrooms. For readers who want to follow the beat or reach the journalists tracking it, here are the correspondents whose work surfaced most often in our corpus.
| Journalist | Publication | Contact | Focus |
|---|---|---|---|
| Andrew Griffin | The Independent | @_andrew_griffin | Debris, reentry pollution |
| Nayara Batschke | The Independent | @nayarabatt | Dark skies, Atacama |
| Samantha Lawler | The Independent | U. Regina profile ↗ | Astronomer voice, space mirrors |
| Aaron Boley | The Independent | aaronboley.com ↗ | Astronomer voice, space mirrors |
| Vishwam Sankaran | The Independent | @Social3uScience | Astronomy, planetary defense |
| Anthony Cuthbertson | The Independent | @ADCuthbertson | Constellations, servicing |
| Ling Xin | South China Morning Post | ling.xin@scmp.com | China debris tracking |
| Stephen Chen | South China Morning Post | binglin.chen@scmp.com | Starlink electronic warfare |
| Victoria Bela | South China Morning Post | victoria.bela@scmp.com | China security framing |
| Angus Dalton | The Sydney Morning Herald | @angus_dalton | Astronomy access, space race |
| Jay Hilotin | Gulf News | — | Starlink milestones, scale |
| Yasmine Alwakal | SBS News | LinkedIn ↗ | Night sky, light pollution |
| TOI Science / Tech Desk | Times of India | @timesofindia | ISRO debris, sat filings |
| Guo Meiping | CGTN | guo.meiping@cgtn.com | Dark-sky, starry economy |
| Nicole Mortillaro | CBC | @NebulousNikki | ISS, Moon & space policy |
| Robert Muggah | Foreign Policy | @robmuggah | Starlink geopolitics |
| Misha Glenny | Foreign Policy | @MishaGlenny | Starlink geopolitics |
Contacts are the authors’ own X handles where available, otherwise a professional profile or newsroom email. The lead Scientific American feature carried no byline in our record; its expert sourcing (Lawler, Lavezzi of MIT, Lifson and Weeden of the Aerospace Corporation, Lewis and Brown of Birmingham, McKnight of LeoLabs, Lemmens of ESA, Pritchard-Kelly) is attributed to the publication.
Ninety seconds. Five mornings a week. Forty tabs, replaced.
The GeoMemo Daily Brief is independent geopolitical intelligence for the people who make the decisions, diplomats, investors, strategists and journalists on deadline. Curated under ICD 203 analytic tradecraft and read across more than 5,000 publications worldwide, it hands you the whole picture rather than one side of it. In your inbox by 08:30, and gone in ninety seconds.
Subscribe free →
© GeoMemo · geomemo.ai