Falcon 9's first stage recovery transformed the economics of launch. SpaceX cut expendable booster costs by recovering and reflying the same hardware dozens of times, and the competitive pressure this created reshaped the entire launch services market within five years. But the first stage is only part of the rocket. The upper stage — which does the majority of the velocity work and delivers payloads to orbit — has remained fully expendable across every operational rocket in history. That's where the next major cost reduction lives, and it is a significantly harder engineering problem.
Why Upper Stages Are Different
A first stage returns from altitudes of 60 to 80 kilometers, having reached peak velocities of 2 to 3 kilometers per second before engine cutoff. Reentry heating is manageable with a combination of controlled descent, grid fins, and a terminal landing burn. The hardware is complex but the physics is tractable.
An upper stage reaches orbital velocity — 7.8 km/s for low Earth orbit, more for higher inclinations and altitudes. Returning from orbit means shedding nearly all of that kinetic energy as heat. The Falcon 9 second stage, if recovered, would need to survive roughly ten times the thermal load of the first stage during reentry, while also being light enough to not significantly penalize payload capacity. Every kilogram of thermal protection system, every actuator for a landing leg, every kilo of residual propellant needed for a deorbit burn — all of it comes directly out of the payload you can sell.
SpaceX ran a paper study on Falcon 9 upper stage recovery in the early 2010s and concluded the mass penalty made it economically unattractive. They moved on to Starship instead.
Starship: The Whole-Vehicle Approach
SpaceX's solution to the upper stage problem was to redesign the entire architecture around a fully reusable second stage large enough that the recovery systems don't consume a prohibitive share of its payload capacity. Starship's Ship (the upper stage) carries roughly 100 metric tons of propellant to accomplish its own landing — but because the vehicle can carry 100-plus tons to LEO when expendable and perhaps 20 to 40 tons reusably, the absolute mass available for recovery hardware is far larger than on a smaller vehicle.
The Ship uses ceramic tiles for reentry thermal protection and performs a belly-flop orientation during descent to maximize drag and spread heating across a larger surface area. Catch and recovery via the launch tower's mechanical arms — the "chopstick" system — eliminates the structural penalty of landing legs. The system has demonstrated it works at the test article level; operational reliability at commercial cadence is the open question.
Rocket Lab's Neutron: Purpose-Built for Reuse
Rocket Lab's Neutron, targeting a first launch in the late 2020s, takes a different architectural position. Rather than going very large like Starship or adapting an existing design, Neutron is a medium-lift rocket designed from the first line of CAD for full reusability. Its upper stage is designed to be recovered — a genuine departure from every other medium-lift vehicle in development.
The company has been deliberately quiet about exactly how Neutron's upper stage reuse works, but the vehicle's architecture — a carbon-composite structure, an aerodynamic fairing that stays with the upper stage, and a targeted payload capacity of 13 tons to LEO (reusable) — suggests a design heavily constrained by the reentry requirement. Rocket Lab CEO Peter Beck has described the second stage as "the hardest part of the whole vehicle."
New Glenn's Honest Approach
Blue Origin's New Glenn launches with an expendable upper stage. The company has not publicly committed to upper stage recovery for New Glenn, making it the most conventionally designed vehicle of the three. Blue Origin's position may reflect a calculated judgment: for now, the market can sustain medium-to-heavy lift with reusable first stages and expendable upper stages, and chasing upper stage recovery before the first stage business is mature is premature risk.
That said, Blue Origin's long-term architecture — gesturing toward a much larger vehicle sometimes called New Armstrong — has not been formally announced with reuse specifications. Upper stage recovery, if pursued, would likely come with the next generation vehicle rather than a retrofit of New Glenn.
Why It Matters
The economics are straightforward in principle. An upper stage for a large launch vehicle costs between $10 million and $40 million depending on the vehicle and production rate. If you can recover and refly it ten times, that amortizes to $1 million to $4 million per flight — a meaningful reduction in a business where launch prices are under competitive pressure from multiple directions.
The less obvious impact is on launch cadence. A manufacturing-limited upper stage is a bottleneck; a reusable fleet is limited by inspection and turnaround time instead. For companies planning frequent launches of large constellations — a business that will define the next decade of commercial space — upper stage reuse could be the difference between a profitable and an unprofitable operation at scale. Whether any of the three approaches actually works in commercial operation is a question the next five years will answer empirically.
