Can a Radar Satellite Survive a Russian Orbital Ambush?

Twenty-Six Metres: A Number That Stops Making Sense

On the morning of October 19, 2026, a Russian military satellite and a commercial radar spacecraft were sharing the sky above your head in a way that should not have been possible. One was ICEYE-X36, operated by a Finnish-American firm and used by the Ukrainian military to locate targets on the ground. The other was Kosmos-2612.

And they were, if analyst Georgy Trishkin's figures are correct, separated by roughly 26 metres.

Hold that number for a moment. Twenty-six metres is about the length of a city bus.

Both objects were travelling at approximately 7.8 kilometres per second. At that speed, the 26-metre gap between them would collapse to nothing in a little over three milliseconds — faster than the human eye can register a camera flash, faster than a nerve signal travels from your fingertip to your brain. A city bus's length, at orbital velocity, is not a safe distance. It is barely a distance at all.

The tracking company LeoLabs had watched the approach unfold across the preceding 24 hours. Their data showed the separation between the two satellites shrinking from 320 kilometres to 4.8 kilometres in a single day — a collapse in proximity that is not drift, not coincidence, and not the background noise of orbital mechanics. Something was steering.

The official confirmed proximity, drawn from independent monitoring, settled around 100 metres. The 26-metre figure is a softer claim, a single analyst's read of the trajectory data. But even the conservative number, even 100 metres at 7.8 kilometres per second, sits in a regime where the word "close" has stopped doing its ordinary work.

This is where the strangeness begins.

How a Russian Military Satellite Actually Sneaks Up on Its Target

Here is the strange part: in orbit, catching something requires slowing down. That feels wrong. Your instinct says accelerate toward the target, as you would on a highway. But orbital mechanics plays by different rules.

Speed determines altitude, and altitude determines how fast you complete each lap. Brake, and you drop lower, completing your orbit faster, gradually creeping up on the spacecraft ahead. Then you fire your engines again to climb back up and match your target. It is counterintuitive, meticulous, and expensive in fuel at every step.

The formal term is Rendezvous and Proximity Operations, or RPO: the deliberate process of matching another spacecraft's orbital plane, altitude, and velocity until two objects essentially share the same small patch of sky. Getting all three coordinates to converge is engineering of the highest order. It took NASA years of Gemini-program practice in the 1960s to make it routine, and even now each manoeuvre burns propellant a satellite will never recover.

The exact sequence of burns that guides a robot arm to repair a failing satellite is indistinguishable, from the outside, from the sequence that guides a grappling claw toward an adversary's asset. Same math, same fuel budget, same orbital choreography. Intent is invisible from a ground telescope.

This is why experts treat one kilometre as a critical threshold: below that distance, an uncoordinated approach leaves almost no margin between an inspection and an ambush. LeoLabs tracked Kosmos-2612 collapsing that gap from 320 kilometres to 4.8 kilometres inside a single day. The kilometre threshold, and everything beyond it, had already been crossed.

A Radar Eye Over a War: Why ICEYE-X36 Was Worth Chasing

To understand why Kosmos-2612 made that burn, you need to know what ICEYE-X36 actually does. It carries a synthetic aperture radar — SAR for short — which means it does not photograph the ground so much as paint it with radio waves and read the echoes back. Cloud cover is irrelevant. Smoke from a burning depot is irrelevant. Time of day is irrelevant.

The Ukrainian military has been using ICEYE imagery to identify hostile targets and coordinate strikes with a precision that simply was not available to any non-superpower army a decade ago.

The practical consequence of that capability shows up in specific, uncomfortable ways for Russia. Satellite imagery revealed that Russia constructed a major launch base for Geran-type drone platforms in roughly six weeks — bases that, once spotted from orbit, became targets. The eye in the sky did not have to fire a single shot; it just had to look, and looking was enough to change the equation on the ground.

This is the wider pattern Zelensky named when he reported widespread throttling of Starlink and Starshield services on the battlefield: the commercial space layer has become a military layer, and everyone involved knows it. Knock out the radar, slow the data link, drain the satellite's fuel with an unannounced orbital visit — these are all moves in the same game.

The target is not the hardware itself so much as the information flow it enables. ICEYE-X36 was worth chasing precisely because it works. That is a new kind of logic for a new kind of war, and it is playing out 500 kilometres above the weather.

The target is not the hardware itself so much as the information flow it enables.

Russia's Inspector Satellites: A Decade of Orbital Rehearsals

What Kosmos-2612 did above Finland in October 2026 was not improvised. It was the latest scene in a decade-long rehearsal, each act a little bolder than the last.

Start with the architecture. Russia's Project Nivelir, developed by the Central Scientific Research Institute of Chemistry and Mechanics (TsNIIKhM), builds satellites that look like one object on a radar screen but are not. Once in orbit, they open and release smaller inspector sub-satellites, a Russian nesting doll trick at 500 kilometres altitude. The outer shell provides cover; the inner unit does the work.

Then came the demonstration. In 2020, a satellite called Kosmos-2543 released a small, fast-moving object in the vicinity of a US satellite. US officials interpreted it as a test of a space-based kinetic weapon, a projectile designed to destroy another spacecraft through sheer collision energy. Nobody called it an act of war. The line between "test" and "threat" remained, deliberately, undrawn.

The intelligence-gathering dimension runs in parallel. The Luch-5X and Olimp-K2 satellites spent years drifting close to Western communications satellites, including Intelsat 39, eavesdropping on their signals. This is called SIGINT stalking, and it is patient work: you park nearby, listen, leave. No collision, no debris. Nothing that fits neatly into any existing treaty violation.

By 2024, the program had grown more direct. US Space Command classified Kosmos-2576, launched in May of that year, as a counterspace weapon, noting it was maneuvering to trail the American reconnaissance satellite USA-314. That same February, White House spokesperson John Kirby warned that Kosmos-2553 represented a Russian effort to develop a space-based nuclear device capable of generating an electromagnetic pulse — a weapon that could blind entire satellite constellations in one detonation.

Each step came with plausible deniability. Each step was also a rung on a ladder leading somewhere very specific.

The Watchers Nobody Hired: How We Know What Happened

Before any government spokesperson had confirmed a single detail about Kosmos-2612's approach to ICEYE-X36, a commercial radar tracking firm in California had already logged the geometry. LeoLabs, which operates a global network of phased-array radar stations, recorded the gap between the two satellites shrinking from 320 kilometres to 4.8 kilometres inside twenty-four hours. No classification level. No press embargo. Just physics, radar pulses, and publicly posted orbital data.

This is the new paradox of orbital transparency. Governments, historically the only parties with the infrastructure to watch objects in low-Earth orbit, now find themselves racing to catch up with commercial tracking networks whose business model depends on telling customers exactly what is happening above their heads. Russia cannot credibly deny a manoeuvre that a private firm in San Jose documented before Moscow's morning briefing was written.

LeoLabs and Orbital Radar publish close-approach data in near real-time, faster and often in greater detail than official Space Command disclosures.

The verification chain is also growing sharper from below. Researchers are optimizing adaptive optics systems to correct for atmospheric scintillation, the shimmering distortion that makes optical observation of low-Earth-orbit objects so difficult from the ground. The same turbulence that blurs a star also scrambles an optical confirmation link tracking a 26-metre close approach. Better optics means better evidence. The watcher's tools are improving faster than the rule book is being written.

An Arms Race With No Rulebook: What Comes Next

In September 2026, the US Space Force confirmed it had placed unspecified weapons in orbit, likely including communication jammers. No further details were given. That is the answer no one wanted to arrive at, and here we are.

The 1967 Outer Space Treaty prohibits weapons of mass destruction in orbit. It says nothing about minimum distances between spacecraft. There is no clause that makes a 26-metre approach illegal, no threshold that converts a non-consensual rendezvous into an act of war. Both sides are currently operating in that silence.

The most insidious tactic may never fire a single projectile. Forcing a satellite into repeated defensive manoeuvres burns finite propellant; run that math enough times and you deorbit a satellite without ever touching it. Georgy Trishkin called it "noncontact deorbiting through fuel depletion." It is attrition conducted at 7.8 kilometres per second, invisible to anyone not watching a tracking screen.

What we still do not know is considerable: Kosmos-2612's actual payload, whether ICEYE-X36 manoeuvred at all, what precisely the US placed in orbit. The open question — the one that each new Russian military satellite manoeuvre sharpens without answering — is where exactly the line sits between a provocation and an act of war. We have no rulebook. We are writing it, badly, in real time, 500 kilometres up.