The Gas Station in the Sky: What's Actually in Orbit

Overview

A satellite in geostationary orbit rarely fails because its hardware breaks. It fails because it runs out of propellant. Cameras, transponders and solar arrays routinely outlast the fuel needed to hold a satellite in its assigned slot, and when the tank empties the spacecraft becomes an expensive piece of debris with functioning equipment aboard.

One answer to that problem has now been demonstrated end to end. Northrop Grumman's Mission Extension Vehicle docked with Intelsat 901 in February 2020, supplied five additional years of service, and undocked in April 2025 — the first undocking between two commercial spacecraft in the geostationary graveyard.

The other answer, actually transferring propellant between spacecraft in orbit, has not happened yet. The first attempt involving a US Department of Defense asset is scheduled for the middle of 2026, and the first commercial fuel depot launches on the same rocket.

These two things are frequently described interchangeably. They are not the same, they involve different engineering, and only one of them has flown.

Why a Working Satellite Dies With a Full Camera

Geostationary satellites are not parked. Solar radiation pressure, lunar and solar gravity, and the Earth's own uneven shape all push a spacecraft out of position, and it must fire thrusters continually to stay in its assigned longitude slot and correct its inclination.

That station-keeping burn rate sets the mission clock. A GEO satellite is typically designed for around fifteen years, and that figure is fundamentally a fuel budget rather than a hardware rating. Electronics and optics regularly remain serviceable well past it.

The result is an unusually wasteful failure mode. An operator retires a satellite that still works, launches a replacement costing hundreds of millions, and pushes the old spacecraft into a graveyard orbit — all because a tank is empty.

What MEV-1 Actually Proved

Northrop Grumman's Space Logistics subsidiary approached the problem sideways. Rather than refuelling Intelsat 901, the Mission Extension Vehicle docked to it and took over the job the propellant was doing — providing attitude and orbit control for the combined stack using its own thrusters and its own fuel.

The docking took place in February 2020, nearly nineteen years after IS-901 launched. The vehicle flew the pair for five years, then undocked in April 2025 after moving the satellite to a graveyard orbit. Intelsat became the first operator to complete a life-extension mission, and IS-901 was decommissioned after twenty-four years of service against a design life of roughly fifteen.

That is a complete, verified commercial result rather than a demonstration. The economics, the docking, the multi-year operation and the clean separation have all been shown to work.

Why Docking Without a Refueling Port Mattered

The most consequential detail of the MEV design is what it did not require. It docks using features the client satellite already has, rather than a purpose-built servicing interface.

That distinction determines the size of the addressable market. A servicing vehicle needing a dedicated port can only ever help satellites launched after that port becomes standard — a market that starts at zero and grows slowly. A vehicle that can grapple existing hardware can service the fleet already in orbit, including spacecraft launched decades before anyone planned for servicing.

The MEV is designed for a fifteen-year life and can dock and undock repeatedly, moving between customers. One vehicle amortised across several clients is what turns the concept from an expensive stunt into a business.

Life Extension Is Not Refueling

This is the distinction that most coverage collapses, and it changes what the technology can do. A mission extension vehicle is a tug. It clamps on and becomes the propulsion system for both spacecraft, and when it eventually runs low the pair are done.

Refueling transfers propellant into the client's own tanks and then departs. The client regains full independent manoeuvrability, and the servicing vehicle moves on with no lasting attachment. It is a fundamentally more flexible architecture and a considerably harder engineering problem, involving fluid transfer of a toxic propellant across a mated interface in vacuum.

Life extension buys time on a fixed platform. Refueling restores capability. Only the first has flown.

The First Real Refueling Attempt Comes in 2026

Astroscale U.S. is targeting the middle of 2026 for a mission that would be several firsts at once: the first hydrazine refueling operation above geostationary orbit, and the first on-orbit refueling of a Department of Defense asset.

The refueler, APS-R, launches with roughly 66 pounds of hydrazine and uses Orbit Fab's refueling interface to mate with its clients. Its targets are Space Force spacecraft from the Tetra-5 programme.

Why the Refueler Refuels Itself

The middle step is the one that matters strategically, and it is easy to overlook as a technicality. A refueler that arrives with a fixed load is a single-use delivery vehicle — useful, but its economics are barely better than launching propellant on demand.

A refueler that can replenish from a depot becomes a reusable shuttle. The expensive, complex rendezvous-and-docking hardware gets amortised across many transfers, while the cheap commodity — the propellant — is delivered in bulk by whatever launch vehicle is most economical.

Orbit Fab's first depot launches on the same rocket as the refueler in June 2026. Demonstrating that handoff is what would separate a one-off servicing mission from the beginnings of an actual distribution network.

What the Contracts Reveal About Confidence

Public contract values are a useful reality check on how seriously the US government treats this, and the numbers are modest by defence-programme standards — the funding profile of a serious experiment rather than an operational capability.

Why the Military Is Driving This

Commercial operators want longer revenue life from expensive assets, which is a real but incremental incentive. The Space Force wants something different: the freedom to manoeuvre without shortening a satellite's life.

In a contested orbital environment, moving is how a satellite avoids being tracked, approached or targeted. Under current constraints every evasive burn is drawn from the same finite budget that keeps the spacecraft on station, so manoeuvring defensively means dying sooner. Delta-v becomes a strategic resource to be hoarded rather than used.

Refueling changes that calculation, which is why the first customer for this technology is a military one and why the demonstration targets are Space Force spacecraft rather than commercial satellites.

What Has to Go Right Before There's a Network

A successful 2026 demonstration would prove the physics and the operations. It would not create an orbital fuel economy, and the remaining obstacles are mostly unglamorous.

Refueling interfaces have to become standard equipment on new satellites, which means operators paying for mass and complexity today against a servicing market that does not yet exist. Depots need enough traffic to justify their own launch costs. And the price per kilogram delivered has to beat the alternative of simply launching a replacement satellite — a bar that keeps moving as launch costs fall.

The honest framing for 2026 is that the tow truck has been proven and the fuel truck is about to attempt its first delivery. Calling that a network of orbital gas stations gets ahead of the evidence by several years.

Conclusion

The disposable satellite is not dead yet, but it now has a demonstrated alternative. Intelsat 901 flew for twenty-four years against a fifteen-year design life because a servicing vehicle took over the job its empty tanks could no longer do, and that mission has run its full course and ended cleanly.

Refueling is the more transformative capability and the less proven one. If APS-R transfers hydrazine to a Space Force satellite in 2026, tops itself up from an Orbit Fab depot, and does it again, the architecture behind a genuine orbital supply chain will have been shown to work in the only environment that counts.

Until then the accurate statement is narrower than the headlines: one satellite has been given extra life by a tug, no satellite has ever been refuelled in orbit, and the first attempt is a few months away.

Data Source and Attribution

AstroscaleNorthrop GrummanSpaceNews

Life-extension figures come from Northrop Grumman and Intelsat's announcements of the IS-901 mission, including the April 2025 undocking. Refueling mission details, propellant loads, launch timing and contract values come from Astroscale U.S., the Defense Innovation Unit, Space Systems Command and contemporaneous reporting by SpaceNews and Air & Space Forces Magazine. Mission dates are targets and remain subject to change.

FactsFigs reviews, cleans, and cross-checks every source dataset before shaping it into a data story. Each visualization is created and designed in FactsFigs Design Studio — an internal tool developed and owned by FactsFigs — and is the original work of a FactsFigs author, not an AI-generated copy of any existing graphic. Individual assets within a visual may or may not be produced with AI tools, but the design of the visual itself is solely FactsFigs' own.

Figures are estimates at the time of publication, provided for information only — nothing here is financial advice or a guarantee of accuracy.

2026-07-20