Jonathan McDowell, an astrophysicist at the Harvard-Smithsonian Center for Astrophysics, publishes space data, launch statistics, and the Jonathan's Space Report newsletter.
Commenters cite it as an authoritative source: GCAT catalogs artificial space objects, and the Starlink statistics page gives satellite counts.
The launch escape system is part of the Crew Dragon spacecraft, which has not entered service yet, and would probably not be used to deliver satellites anyway because its crew-carrying capabilities add significant weight and cost.
Rocket launches have gotten fairly reliable in recent years -- launch success rates have been hovering around ~95% since the 1970s. [1][2] A 5% chance of killing 7 astronauts is unacceptably high at essentially any cost, hence the launch escape system on the Crew Dragon. However, for an unmanned launch, the entire launch escape and recovery system would have to add little more than 5% extra cost on top of the satellite and launch vehicle for it to be a sound purchase.
A Falcon 9 rocket costs about $62M. A very complex satellite can cost over $200M, so for easy math, let's say that the entire system costs $300M, 5% of which is $15M. Those $15M have to cover the following:
- powerful launch escape engines (e.g. SuperDracos) that are not recovered if the satellite reaches orbit
- hydrazine fuel for those engines (highly toxic and reactive, has to be handled with great care)
- carrying all that extra weight to orbit
- parachutes and their deployment systems
- hardening the payload to survive the physical stresses experienced during landing
- refurbishing the payload after it falls into corrosive salt water
Given all this, I am skeptical that a launch escape system for unmanned payloads is economically feasible.
[1]…
Starlink isn't at the top of my list for "likely Kessler Syndrome starters" because its a rather low orbit.
They're at 550km up. From https://www.spaceacademy.net.au/watch/debris/orblife.htm puts it at likely a 10 year lifespan with some variation. The published lifespan of a satellite is 5 years.
Its the stuff in the 700km to 900km range that's a problem.
http://stuffin.space
https://www.satview.org/spacejunk.php - and you'll see things like STARLINK-1204 already getting deorbtied ( https://www.n2yo.com/satellite/?s=45203 - note its perigee is 204.0 km - its already well below the other orbits)
https://in-the-sky.org/spacecraft.php?id=45203 for more on that satellite.
Starlink-61 appears to be one of the oldest ones still in orbit - https://in-the-sky.org/spacecraft.php?id=44249 though its on its way down and you can see the type of impact and the duration of the "ok, this is going down".
Even with the maximum starlink, that's 1,700 satellites that will last about 10 years.
The 2007 China anti-satellite test created approximately 150,000 debris at 865km. Those are in an orbit that is on the order of 500 years.
Starlink isn't going to be the cause of Kessler Syndrome.
---
In hunting this up, I also found https://planet4589.org/astro/starsim/index.html which is relevant to the main topic - the impact on astronomy.
A dramatic factor within itself, not a dramatic quantity compared to everything else.
I'm not sure exactly how much aviation fuel China burned in 2025, nor mass to orbit, but a quick search suggests that in 2024 they burned 38.2 million tonnes of aviation fuel[0], and launched 174.4 tonnes to orbit[1]. For the sake of a Fermi approximation, round that up to 40 MT aviation emissions and 200 tonnes to orbit, and assume a 20x factor of propellant CO2 emission mass to payload mass:
200 * 20 = 4e3 tonnes of CO2 from space launches
40e6 / 4e3 = 10,000 times more comes from aviation
So, space could go up by a factor of 100x (compared to 2024) and still be a literal rounding error compared to just aviation.
As per article, graph shows the nation as a whole is just over 12,000 million tonnes "from fossil fuels and cement".
12e9 / 4e3 = 3,000,000
If I download the image to get the full resolution, there's 586 pixels between the lines for 10,000 million and 12,000 million tonnes, so each pixel on that graph represents 2e9 / 586 ~= 3.4 million tonnes, and 3.4e6 / 4e3 = 850, so even if space launch grew by a factor of a thousand it would only shift that graph by one pixel. One and a bit, but closer to one.
[0] Search for "38.2": https://www.greenairnews.com/?p=7427
[1] Page 13: https://planet4589.org/space/papers/space24.1.0.pdf
The article says "Instead, the spaceman (and only a man was considered) would have been launched on a parabolic trajectory some 300,000 metres above the Earth."
I think that's a mistake. The V2 had a range of about 300 km ("For the V-2 rocket, just reaching space but with a range of about 330 km, the maximum speed was 1.6 km/s" - https://en.wikipedia.org/wiki/Sub-orbital_spaceflight#Speed.2C_range.2C_altitude ).
The BBC article goes on to quote “Megaroc was essentially the Mercury-Redstone.” From https://en.wikipedia.org/wiki/Mercury-Redstone_Launch_Vehicle we see that Mercury-Redstone didn't quite make it to 200 km altitude.
According to http://www.planet4589.org/space/jsr/back/news.526 , there have been no suborbital flights above 192 km.
EDIT: Nope, I'm wrong. Megaroc was supposed to get to 304 km (one million feet). http://www.bis-space.com/what-we-do/projects/megaroc . Indeed, it appears that the name is short for "Mega[foot]roc[et]."
Not feasible. That would entail putting shrapnel into orbit (unlike extant anti-sat weapons which are short-range suborbital), which would mean a full orbital launch for every satellite target orbit. There's hundreds[0] of Starlink orbital groups already, so that'd require hundreds of independent orbital launches in a short timescale—far beyond China's launch capabilities today.
[0] https://planet4589.org/space/con/star/planes.html
(On general principles, you could argue you'd need 1:1 launch vehicle parity (number, not payload) to defeat a satellite constellation this way. For each satellite launch, you'd need one corresponding anti-satellite launch into that same, newly-defined orbit).
>absorbent or light diffusing material should be mandatory
Absorbency is bad because it causes overheating and IR astronomy pollution. Diffusion is bad because it reflects light toward Earth.
Specular surfaces (mirrors) can reflect light entirely into space and away from observers on the ground.
This explains why SpaceX created their own radio-transparent, space-rated mirror. They are offering it—at cost—for use by other satellite manufacturers.
https://www.teslarati.com/spacex-shares-how-its-making-starlink-satellite-less-bright/
https://planet4589.org/astro/starsim/2022-05-07_SpaceX_Sustainability_for_FAS_V3.pdf