Sending astronauts to the Moon is difficult. Sending an enormous spacecraft there after filling it with fuel through a series of orbital transfers could be one of the most complicated engineering operations ever attempted. SpaceX’s Starship was designed to make humanity a multiplanetary species, but an unexpected challenge is emerging: before it can carry astronauts to the lunar surface, it may need an entire fleet of rockets working together in space.
For more than half a century, the Moon has remained a destination that humanity has struggled to revisit.
NASA’s Apollo programme put 12 astronauts on the lunar surface between 1969 and 1972. Those missions relied on the powerful Saturn V rocket, which carried the necessary fuel from Earth in a single launch.
SpaceX wants to approach the problem differently.
Its enormous Starship spacecraft is designed to be reusable, carry substantial cargo and eventually transport astronauts to the Moon and Mars. But there is a catch: Starship needs an extraordinary amount of propellant to complete a lunar mission.
Rather than launching with all its fuel, the spacecraft is expected to receive propellant in Earth orbit from specially configured tanker vehicles.
And according to a recent estimate discussed by a member of NASA’s Aerospace Safety Advisory Panel, the lunar mission architecture could require approximately 22 Starship flights to support a single Moon landing.
That figure has attracted considerable attention because it illustrates just how dependent the next generation of lunar exploration could become on orbital refuelling.
There is an important qualification, however: 22 is an estimate, not a final NASA requirement. SpaceX CEO Elon Musk has publicly disputed it, suggesting the number could eventually be closer to six or eight.
The disagreement reveals how much engineering work remains before Starship can routinely transport astronauts beyond Earth.
Why can’t Starship simply carry all its fuel from Earth?
The answer comes down to a fundamental problem in rocket engineering: mass.
A rocket needs fuel to accelerate itself, its payload and the fuel it will consume later in the journey.
But the more fuel it carries, the heavier it becomes. That additional mass requires even more fuel to move, creating a difficult engineering equation known as the rocket equation.
Starship is designed to carry substantial quantities of liquid oxygen and liquid methane, which power its Raptor engines.
These propellants are attractive because methane can be produced on Mars using locally available resources, at least in principle. They also provide the performance needed for a reusable spacecraft.
However, transporting enough propellant for a lunar journey, landing and subsequent return requires an enormous amount of energy.
SpaceX’s proposed solution is to separate the transportation of fuel from the transportation of astronauts.
Instead of launching one fully fuelled lunar spacecraft, it can launch multiple tankers that gradually build up the required propellant supply in orbit.
This is called orbital propellant aggregation.
The concept could eventually transform how large spacecraft are designed. Instead of treating every launch as a self-contained journey, rockets could become part of a larger transportation network operating hundreds of kilometres above Earth.
How the 22-refuelling estimate emerged
The number that triggered the recent discussion came from Charlie Precourt, a former NASA astronaut and member of the Aerospace Safety Advisory Panel.
During the panel’s October 2026 meeting, Precourt discussed the risks associated with NASA’s Artemis IV programme and said the latest estimate for the required Starship flights was around 22.
The figure was subsequently reported by Ars Technica on October 22.
Elon Musk responded publicly, arguing that the estimate was too high and suggesting that the actual requirement could be closer to six to eight flights.
Why such a large difference?
The answer involves several variables, including Starship’s payload capacity, the amount of propellant needed, fuel losses during storage, transfer efficiency and the final mission architecture.
NASA’s earlier planning documents also discussed the need for more than 10 tanker flights, alongside a dedicated propellant storage depot.
The important distinction is that 22 refers to the estimated launch activity supporting the lunar landing architecture, not 22 separate refuelling operations performed by the astronauts’ spacecraft.
Read the original reporting from Ars Technica.
The orbital fuel depot: A petrol station hundreds of kilometres above Earth
The proposed refuelling system is essentially a space-based fuel station.
The process is expected to involve three different types of Starship configurations:
1. The tanker spacecraft
These Starships will carry additional liquid oxygen and liquid methane into low Earth orbit. Their job is to deliver fuel rather than transport astronauts to the Moon.
2. The propellant depot
This specialised Starship will act as a storage facility. It will receive fuel from arriving tankers and hold the accumulated propellant until the lunar lander is ready.
3. The lunar lander
The human Landing System version of Starship will receive the stored propellant and use it to travel towards the Moon, descend to the lunar surface and eventually return its crew to lunar orbit.
The architecture is ambitious because these spacecraft must rendezvous, dock and transfer enormous quantities of extremely cold liquids while travelling at orbital speeds.
NASA’s Human Landing System programme provides an overview of the technologies required to support astronaut transportation between lunar orbit and the surface.
Why transferring fuel in space is so difficult
On Earth, refuelling a vehicle is straightforward. A pump moves liquid through a hose, gravity helps keep the liquid where it belongs, and the process can be monitored using familiar equipment.
In orbit, the physics is considerably more complicated.
First, there is microgravity. Liquids do not naturally settle at the bottom of tanks in the way they do on Earth. They can form floating masses, cling to surfaces or move unpredictably.
Engineers must develop systems that control liquid movement and ensure the propellant reaches the correct transfer connections.
Second, there is temperature.
Starship uses liquid oxygen and liquid methane, which must be maintained at extremely low temperatures. Liquid oxygen is stored at approximately minus 183°C, while liquid methane is stored at around minus 162°C under typical storage conditions.
Keeping these propellants cold in space is a major engineering challenge.
Heat gradually enters storage tanks from surrounding structures and equipment. Even relatively small amounts of warming can cause some liquid to evaporate, a process called boil-off.
If too much propellant turns into gas during a long refuelling campaign, the amount available for the lunar journey decreases.
Finally, there is the challenge of docking.
Two enormous spacecraft must approach one another at carefully controlled speeds, align their docking systems and establish a secure connection before transferring fuel.
A failure at any stage could jeopardise the entire operation.
SpaceX has already demonstrated a small-scale milestone: its March 2025 Starship flight test transferred approximately five metric tonnes of liquid oxygen between tanks aboard the same vehicle. However, this was an internal transfer demonstration, not a full-scale ship-to-ship orbital refuelling operation.
The company still needs to demonstrate the more demanding operation involving separate spacecraft.
SpaceX’s Starship development updates describe its progress towards orbital propellant transfer.
NASA has already seen a glimpse of the technology
The idea of orbital refuelling is no longer entirely theoretical.
During a Starship flight test in March 2025, SpaceX demonstrated the transfer of cryogenic propellant between tanks inside the same spacecraft.
Although this was an important engineering achievement, transferring fuel between separate spacecraft introduces additional challenges.
NASA wants SpaceX to demonstrate the complete process before relying on it for a crewed lunar mission.
The US space agency’s Inspector General has also identified orbital refuelling as one of the most significant technical challenges facing the Starship programme.
In its March 2026 assessment, NASA’s watchdog highlighted concerns about the maturity of cryogenic storage and transfer technologies, as well as the programme’s development schedule.
Read NASA’s Office of Inspector General report.
NASA has changed the Moon mission architecture
There is another important development that makes the 22-flight estimate particularly interesting.
On October 23, 2026, NASA announced changes to its Artemis lunar landing architecture.
Under the revised approach, Starship will collect propellant in low Earth orbit before travelling towards the Moon. The spacecraft will then rendezvous with NASA’s Orion spacecraft in Earth orbit, rather than following the earlier planned sequence involving a rendezvous in lunar orbit.
NASA says the revised approach reduces the amount of propellant Starship needs to aggregate in space.
This could influence the number of tanker launches required.
However, NASA has not published a final confirmed number that establishes 22 launches as the requirement under the updated architecture.
The revised mission also changes the planned Artemis III objectives, with a low-Earth-orbit demonstration of rendezvous and docking between the lander systems and Orion.
Read NASA’s October 2026 Artemis announcement.
What happens if orbital refuelling succeeds?
The consequences could extend far beyond NASA’s immediate objective of returning astronauts to the Moon.
If SpaceX successfully demonstrates reliable orbital refuelling, it could fundamentally change the economics and engineering of deep-space transportation.
Large spacecraft would no longer need to carry every kilogram of fuel from Earth’s surface.
Instead, reusable tanker vehicles could supply propellant to spacecraft already waiting in orbit.
This could support:
-
Larger lunar exploration vehicles.
-
Delivery of heavy equipment and lunar habitats.
-
More ambitious robotic missions.
-
Future Mars transportation systems.
-
Potentially reusable deep-space cargo vehicles.
For Mars exploration, the technology could be particularly important. A spacecraft travelling to the Red Planet would require substantial propellant reserves, especially if the mission includes landing, surface operations and eventual return.
However, Mars missions would introduce additional challenges, including longer-duration cryogenic storage and the need to protect crews during extended journeys.
Orbital refuelling is therefore not simply a solution for reaching the Moon. It could become one of the foundational technologies for human exploration beyond Earth.
The Moon is waiting, but the engineering must come first
There is something extraordinary about the possibility of launching more than a dozen rockets simply to fill another rocket with fuel.
It sounds inefficient compared with the Apollo missions, which sent astronauts to the Moon using a single Saturn V launch.
But Starship is being designed around a different philosophy: reusable spacecraft, high launch frequency and the ability to transport much larger quantities of cargo.
The 22-flight estimate highlights the scale of the challenge, not necessarily the final number of launches that will be required.
SpaceX still needs to demonstrate full-scale ship-to-ship cryogenic propellant transfer, reliable long-duration fuel storage and the safety systems needed for human lunar missions.
NASA must also verify that the complete system can meet its astronaut safety requirements.
If these challenges are overcome, the Moon could become much more accessible than it has been since the Apollo era.
And perhaps the most remarkable part of this story is that the next giant leap for humanity may not depend on building a more powerful rocket alone.
It may depend on learning how to turn Earth’s orbit into a giant refuelling station, where spacecraft meet, exchange fuel and prepare for journeys to other worlds.
