← Daily DoseEverything We Left Up There — Daily Dose No. 003
Daily Dose / Aug 5, 2026 / Orbital Debris
Everything We Left Up There
Seventy years of spaceflight has put roughly 17,000 tonnes of hardware into orbit around Earth — and left most of it there. Some of it still works. Far more of it is shrapnel, travelling at ten kilometres a second through the same thin shells of space we depend on for weather, navigation, and the internet.
141million fragments
…larger than one millimetre are estimated to be orbiting Earth right now. Surveillance networks can actually see about 46,120 of them. Everything else is a bullet nobody is tracking.
ESA MASTER model & DISCOS catalogue · 31 Jul 2026
01 / The size spectrum
A cloud you mostly cannot see
Debris population is inversely stacked against our ability to observe it: the smaller the fragment, the more of them there are — and the harder they are to catalogue. The lethal middle band is the problem.
1 mm — untrackable, erosive1 cm — lethal, mostly invisible10 cm+ — catalogued
1 mm – 1 cm
140,000,000
Paint flecks, insulation, solid-motor slag. Pits windows and strips solar cells. Effectively untrackable.
1 cm – 10 cm
1,200,000
Too small to reliably track, big enough to end a mission. The single most dangerous class in orbit.
Larger than 10 cm
54,000
Bolts, panels, dead satellites, spent stages. Mostly catalogued — and mostly what gets dodged.
Source: ESA Space Environment Report 2026 (MASTER model estimates) — figures are modelled, not counted
Impact energy · why 1 cm matters
Average collision speed in low Earth orbit is around 10 km/s — roughly 36,000 km/h. At that speed a 1 cm aluminium sphere weighing 1.4 grams arrives with about 71 kilojoules of kinetic energy: the same wallop as a 1,500 kg car hitting a wall at 35 km/h. There is no armour that is practical at that scale. [computed from standard LEO closing speed — illustrative]
02 / The inventory
What is actually up there
Of the 27,490 satellites humanity has launched since Sputnik, about 18,840 remain in space. Only a fraction still work. The catalogue of currently tracked objects breaks down like this.
Source: ESA DISCOS database, 31 July 2026 · “unidentified” = tracked but not yet attributed to a launch
Total mass in orbit
17,000 t
More than the steel in the Eiffel Tower, twice over — and still climbing by roughly 2,000 tonnes a year.
Satellites launched
27,490
Since 1957, across about 7,320 successful rocket launches.
Still active payloads
16,234
Working spacecraft as of 8 July 2026. Around two thirds of them are Starlink.
03 / Origins
Six hundred and sixty-one break-ups
Most debris was not left behind — it was made, in an instant, by an explosion or a collision. Since 1961 there have been 661 confirmed in-orbit fragmentations, 207 of them in the last twenty years. Three events did outsized damage.
Sources: NASA History of On-orbit Satellite Fragmentations (16th ed.); ESA Space Environment Report 2026
Accidental fragmentations
10.5/yr
The long-run annual average of unintended break-ups — leftover fuel and batteries exploding in dead hardware.
Added in 2024 alone
3,000+
Newly catalogued fragments from unintended events in a single year.
Still up from 2007
2,800+
Fengyun-1C fragments still in orbit nineteen years after the test that made them.
04 / The dodging
Avoidance has gone exponential
SpaceX files a collision-avoidance count with the FCC every six months. In three years the number has grown roughly eightfold — partly because there are more satellites, partly because SpaceX now manoeuvres at a risk threshold 300× more cautious than the industry norm.
Source: SpaceX semi-annual constellation reports to the U.S. FCC, 2023–2026
12 months to May 2026
355,848
Total Starlink avoidance manoeuvres — more than three times the 2024 figure.
Manoeuvre threshold
3 in 10M
SpaceX moves at a 3×10⁻⁷ collision probability. The industry standard is 1 in 10,000.
ISS avoidance burns
41
Times the International Space Station has fired thrusters to dodge debris since 1999.
05 / The geography of orbit
Where the crowding is
Orbital space is not uniformly full. Two thin shells carry most of the risk — the mega-constellation band at 500–600 km, and the old sun-synchronous corridor at 800–900 km where decades of Earth-observation hardware and its wreckage still circulate.
Sources: ESA Space Environment Report 2025 & 2026; DLR Space Debris briefing · diagram is schematic
06 / The stakes
Coming down, and going up
Debris leaves orbit two ways: it burns up, or somebody pays to fetch it. Right now the burning is happening more than three times a day — and the fetching has never been done at all.
Collision risk in LEO
+20%
Increase in modelled low-Earth-orbit collision risk since 2024, per ESA's 2026 assessment.
Intact objects re-entering
3+/day
Whole satellites and rocket stages falling back to Earth — some 100–120 tonnes a year uncontrolled.
Cost to remove one object
€86M
ESA's contract with ClearSpace to de-orbit a single Vega adapter. Nothing has been removed yet.
Source: ESA Space Environment Report 2025/2026 — post-mission disposal compliance for LEO rocket bodies
Why it matters
The problem is not the junk. It's the arithmetic.
ESA's blunt finding is that even if every launch stopped tomorrow, fragmentation events would still add debris faster than the atmosphere drags it away. That is the definition of a self-sustaining cascade — the Kessler syndrome Donald Kessler described in 1978. It does not arrive as an explosion. It arrives as a slow closing of altitudes: first 800 km becomes uninsurable, then 550 km, then launch windows start being scheduled around wreckage.
The encouraging part is that this is a compliance problem, not a physics problem, and compliance is improving fast.
Passivation works. Venting leftover fuel and discharging batteries at end of mission eliminates the explosions that cause most accidental break-ups. It costs almost nothing.
The five-year rule is being adopted. About 80% of LEO rocket bodies now meet ESA's 2023 standard of clearing orbit within five years — up from a 25-year norm.
Deliberate fragmentation has to end. Three anti-satellite tests produced over 7,000 trackable fragments. More than 155 nations have now backed a moratorium on destructive ASAT testing.
Watch the sky yourself. Public catalogues let anyone track re-entries and conjunctions in near real time — the data underlying every number on this page is open.
Sources & notes
ESA Space Debris User Portal / DISCOS database, statistics snapshot 31 July 2026 — 46,120 tracked objects; 27,490 satellites launched; ~18,840 still in space; >17,000 tonnes in orbit; catalogue breakdown by object class; >660 break-up events.
ESA Space Environment Report 2026 (Space Debris Office, ESOC Darmstadt) — MASTER model estimates of 54,000 objects >10 cm, 1.2 million at 1–10 cm, 140 million at 1 mm–1 cm; +20% LEO collision risk since 2024; debris density parity with active satellites at ~550 km; average 10.5 accidental fragmentations per year. Population figures below 10 cm are MODELLED, not observed.
ESA Space Environment Report 2025 — intact objects re-entering more than three times per day; ~90% of LEO rocket bodies compliant with the 25-year disposal rule and ~80% with ESA's 5-year rule; >3,000 tracked objects added by 2024 fragmentation events.
NASA, History of On-orbit Satellite Fragmentations, 16th edition (NASA/TP-20220019160) — fragment counts for Fengyun-1C (2007), Iridium 33 / Cosmos 2251 (2009) and Cosmos 1408 (2021).
SpaceX semi-annual constellation reports to the U.S. FCC, 2023–2026 — 207,152 avoidance manoeuvres Dec 2025–May 2026; 148,696 in the preceding six months; 144,404 for Dec 2024–May 2025; ~50,000 for Dec 2023–May 2024; ~25,000 for Dec 2022–May 2023; 3×10⁻⁷ manoeuvre threshold.
NASA Orbital Debris Program Office / NASA ISS records, 2025 — 41 ISS collision-avoidance manoeuvres since 1999, most recently 30 April 2025.
Jonathan McDowell, Jonathan's Space Report, 8 July 2026 — 16,234 active payloads in orbit.
ESA press release, "ESA purchases world-first debris removal mission from a start-up", 2020 — €86 million ClearSpace-1 contract to de-orbit a single Vega secondary payload adapter.
D. J. Kessler & B. G. Cour-Palais, Collision Frequency of Artificial Satellites: The Creation of a Debris Belt, Journal of Geophysical Research, 1978 — the original cascade hypothesis.
DLR Space Debris briefing — peak debris density at 800–900 km altitude in sun-synchronous orbit.
COMPUTED / ILLUSTRATIVE: the 71 kJ impact energy figure is our own calculation for a 1 cm aluminium sphere (1.41 g, ρ = 2.7 g/cm³) at a 10 km/s closing speed, and the car comparison follows from it. Not a measured value.