Satellites are expensive. A single geostationary communications satellite can cost anywhere from $200 million to over $400 million to build and launch. Once it's up there, running low on propellant or suffering a correctable mechanical failure is effectively a death sentence. The satellite drifts, gets decommissioned, and eventually becomes another piece of orbital debris. A replacement gets ordered, built over three to five years, and launched at enormous cost. This has been the accepted model for decades, mostly because no alternative existed.
That's changing fast. On-orbit servicing, the practice of sending spacecraft to repair, refuel, reposition, or upgrade other spacecraft, is shifting from a niche research ambition to something that looks increasingly like a real industry. Several missions are now operating or entering operational timelines in 2024 and 2025, and the stakes are genuinely high. Success means a fundamental rethinking of satellite economics. Failure, or worse, a collision in a crowded orbital band, could set the entire field back by years and trigger a cascade of regulatory headaches that nobody is ready for.
How We Got Here
The idea of servicing satellites in orbit isn't new. NASA astronauts famously repaired the Hubble Space Telescope five separate times between 1993 and 2009, correcting a flawed mirror and upgrading instruments across multiple servicing missions. Those missions demonstrated unambiguously that on-orbit intervention works. But they required human crews, the Space Shuttle, and enormous cost, none of which scaled to commercial satellites.
The first credible commercial push came from Northrop Grumman's Mission Extension Vehicle program. MEV-1 docked with Intelsat 901 in February 2020, becoming the first commercial spacecraft to dock with another commercial satellite. The technique was elegantly simple: MEV-1 physically latched onto IS-901's apogee kick motor nozzle and used its own thrusters to keep the aging satellite on station. It worked. IS-901 got five more years of operational life. MEV-2 followed in 2021 with Intelsat 10-02. Two missions, two successes, and suddenly the industry had proof of concept at commercial scale.
But propellant transfer, in-space refueling rather than just mechanical life extension, proved harder. Northrop's Mission Robotic Vehicle, intended to perform more sophisticated servicing including refueling, has faced development delays. Meanwhile, a new wave of companies has entered the space, each with different approaches and different target markets.
The Current Field of Players
Astroscale has been the most aggressive in demonstrating proximity operations. Their ELSA-d mission tested capture and release of a debris object using magnetic docking technology. Their follow-on ELSA-M is designed for multi-client debris removal in low Earth orbit. The company has also secured contracts with JAXA for active debris removal of defunct JAXA rocket bodies, actual operational missions rather than demonstrations. Their commercial refueling concept, LEXI, is in development targeting GEO satellites.
ClearSpace, backed by the European Space Agency, is targeting a specific piece of Swiss debris for removal with their ClearSpace-1 mission. Originally planned for 2025, the timeline has slipped, but the mission design is technically ambitious: a four-armed robotic spacecraft that will capture a Vega rocket adapter with a claw-style mechanism, then deorbit with it.
In the United States, Momentus has had a rougher road, facing regulatory hurdles and technical setbacks with their water plasma propulsion system. Orbit Fab has taken a different approach entirely, focusing not on performing the servicing but on building the refueling infrastructure. Their "gas stations in space" concept involves pre-positioned propellant depots, and they've signed agreements with multiple satellite operators and servicer companies. Their Rapidly Attachable Fluid Transfer Interface, or RAFTI, is being designed as an industry standard port that future satellites could incorporate at the factory, making refueling straightforward rather than requiring specialized docking hardware.
DARPA has skin in the game too. The agency's RSGS program (Robotic Servicing of Geosynchronous Satellites) is a public-private partnership with Maxar Technologies to develop a robotic servicing vehicle capable of inspecting, repositioning, and repairing GEO satellites. The GEO belt, sitting at 35,786 kilometers altitude, is home to some of the most valuable communications and weather satellites on Earth, and it's where the economic case for servicing is strongest.
Why GEO Is the Prize
Geostationary satellites are worth servicing because they're expensive, they're stationary relative to the ground (making rendezvous predictable), and their operators know exactly when propellant will run out. A typical GEO communications satellite has a design life of 15 years. The primary life-limiter isn't the electronics or the solar panels. It's fuel. Most are designed with just enough propellant to maintain their orbital slot and handle occasional corrections over that period. When the tank approaches empty, the satellite gets moved to a "graveyard orbit" a few hundred kilometers above the GEO belt, and that's that.
Refueling a GEO satellite that still has functioning hardware could extend its life by five to ten years. At $300 million per replacement satellite, that extension is worth hundreds of millions of dollars in deferred capital expenditure for an operator. Even if refueling costs $50 million to $100 million per mission, the math works convincingly. That's why every major satellite operator has been watching the MEV missions with intense interest.
There's also the matter of orbital slots. The GEO belt has finite capacity. The International Telecommunication Union manages slot assignments, and prime slots over densely populated regions are genuinely scarce. A satellite that runs out of propellant and vacates a prime slot creates a competitive scramble among operators to claim it. Extending operational lifetimes keeps valuable assets producing revenue and keeps operators in their hard-won orbital positions.
The Technical and Regulatory Minefield
Getting close to a satellite you don't own is harder than it sounds. Most existing GEO satellites were never designed to be serviced. They lack standardized docking ports. Their propellant systems use hydrazine, a toxic monopropellant that requires careful handling. Their thermal profiles create unpredictable environments for an approaching servicer spacecraft. And critically, if a proximity operation goes wrong and two spacecraft collide in GEO, the resulting debris field could threaten dozens of neighboring satellites in one of the most economically critical orbital regimes on Earth.
Regulatory frameworks haven't kept pace with the technology. In the United States, the FCC, FAA, and NOAA all have overlapping jurisdiction over different aspects of commercial space operations. There's no single, clear regulatory pathway for a company that wants to approach, dock with, and service a third-party satellite. The current rules were written for an era when satellites were launched and left alone. The question of liability, specifically who is responsible if a servicing mission causes damage to a client satellite or a neighboring spacecraft, remains genuinely unsettled in international law.
The Outer Space Treaty of 1967 establishes that launching states bear responsibility for their spacecraft. But a servicing mission involves at minimum two spacecraft, potentially from two different countries, interacting in orbit. The legal ambiguity is real and unresolved, and it's one reason why early commercial servicing agreements have been structured carefully to limit liability exposure.
What Comes After Refueling
Refueling is the entry-level service. The more sophisticated vision involves actual in-space manufacturing and upgrade capability. NASA's OSAM-1 mission (On-orbit Servicing, Assembly, and Manufacturing) was intended to demonstrate robotic refueling of a satellite not designed for it, along with manufacturing a usable beam in orbit from raw material. The project has faced significant cost overruns and schedule delays, ballooning from an initial estimate of around $700 million to over $2 billion before NASA canceled it in 2024. The cancellation stung, but it also reflected a harder-nosed accounting of what taxpayer-funded demonstration missions can reasonably accomplish compared to what commercial ventures are already doing.
Commercial players are now picking up where government programs stall. Starfish Space, a small startup, is developing the Otter spacecraft for in-orbit docking and servicing with a much leaner budget. Their approach uses vision-based navigation and a relatively gentle capture mechanism. They completed a successful proximity operations demonstration with their Otter Pup pathfinder mission in 2024, getting within meters of a target object.
The Bigger Picture for Space Infrastructure
On-orbit servicing is really a proxy for a larger question: can humanity start treating space as a place where things are built, maintained, and operated over long timescales, rather than a one-way destination for hardware that gets launched and eventually abandoned? The answer has enormous implications for the economics of space, the growth of the satellite industry, and eventually for more ambitious programs like lunar infrastructure and deep-space exploration.
For satellite operators and their customers, which ultimately includes anyone using GPS, satellite internet, weather forecasting, or global communications, successful servicing means lower long-term costs and more reliable service. For space enthusiasts watching this field, the next 18 to 24 months will be the most informative period yet. Multiple missions will attempt proximity operations, docking, and propellant transfer in real operational conditions. The results will tell us how close the industry actually is to making in-space servicing as routine as terrestrial maintenance. And the answer to that question will shape the orbital economy for the rest of this century.
