How Long Do Starlink Satellites Last?

Roughly five years is the design life of a Starlink satellite. For anyone used to conventional spacecraft economics, that number looks alarming: enormous capital cost, very short return period.

It makes sense once you notice that SpaceX is not really in the same business.

What actually limits the life

Several things wear out, and the binding constraint is usually propellant.

At 550 km there is still a trace of atmosphere. It is extraordinarily thin, but a satellite travelling through it at 7.6 km/s feels measurable drag, and drag lowers the orbit. Left alone, a Starlink satellite would spiral down and re-enter within a few years. Staying up requires periodic thruster firings, and every firing consumes propellant that cannot be replaced.

When the propellant runs out, the satellite can no longer maintain altitude — and, more importantly, can no longer manoeuvre to avoid other objects. That is the end of its useful life regardless of whether anything has broken.

• Propellant — the usual limiting factor; no refuelling exists for these satellites
• Solar array degradation — radiation slowly reduces power output, and less power means less usable capacity
• Battery cycling — roughly sixteen charge/discharge cycles a day, thousands per year, gradually reducing capacity
• Radiation damage to electronics — cumulative, and eventually causes component failures
• Reaction wheel and mechanism wear — the moving parts that control orientation have finite lifetimes

Why a short life is deliberate

The conventional approach is to build a spacecraft as reliable and long-lived as possible, because launching it is the expensive part. That logic inverts when you own the rocket and launch weekly.

A satellite designed for fifteen years must use technology qualified at design time and then live with it for a decade and a half. In a field moving as fast as phased-array electronics and optical communications, that means the constellation would be permanently a generation or two behind.

A five-year life means the whole constellation refreshes continuously. Newer, more capable satellites replace older ones as a matter of routine, and capacity grows without any special upgrade programme. The V2 Mini generation carries several times the bandwidth of the v1.0 satellites it is gradually replacing — a gain that would have been impossible with fifteen-year hardware.

Short design life is a feature of vertical integration. It only makes economic sense if launch is cheap, and launch is cheap because SpaceX is also the launch provider.

The debris argument

There is a second, genuinely important reason for the low orbit and short life: it makes the constellation self-cleaning.

At 550 km, atmospheric drag removes a dead satellite within about five years even if nothing is done deliberately. Compare that with the graveyard problem in higher orbits — an object abandoned at 1,200 km stays there for centuries, and one in geostationary orbit is effectively permanent.

This matters because low Earth orbit's worst-case scenario is a cascade of collisions generating debris faster than it decays. An orbit that clears itself on a five-year timescale is structurally resistant to that. It is one of the strongest arguments in Starlink's favour on space-sustainability grounds, and it is a direct consequence of choosing an altitude where the atmosphere still does work for you.

What happens at end of life

A healthy satellite reaching end of life is deorbited deliberately. It uses its remaining propellant to lower its orbit, drag takes over, and it re-enters and burns up. SpaceX aims for full demise — the satellite is designed so that no component survives to reach the ground.

A satellite that fails early is a different case. This is why the deployment strategy matters: new satellites are released into a low parking orbit, only a few hundred kilometres up, and must raise themselves. A satellite that fails its checkout never leaves that low orbit and re-enters naturally within months. Failures are removed by design rather than by intervention.

What this means for launch cadence

The arithmetic is unforgiving. A constellation of several thousand satellites with a five-year life needs to replace roughly a fifth of itself every year, forever, just to stay the same size.

That is why Starlink launches are so frequent and why they will not stop. A large share of every year's launches is replacement rather than growth. It is also why SpaceX's launch cost per kilogram is not a side detail — it is the number the entire business model rests on, and why Starship's economics matter so much to the constellation's future.

Watching it happen

The turnover is visible in the data. The statistics page on this site groups the current constellation by launch year, and the shape of that distribution shows the replacement cycle directly — early years thinning out as those satellites are retired, recent years dominating the active count.

Individual satellites tell the same story. A satellite whose altitude is drifting downward over weeks rather than holding steady is either low on propellant or already on its way out.