Satellite Service Stations: The Race to Build On-Orbit Logistics for Defence Constellations

We build satellites. We launch them into orbits worth hundreds of millions of dollars. And when they run out of fuel, develop a fault, or simply age out of their intended position — we leave them there. No service center. No tow truck. No mechanic.

That paradox — the most expensive machines ever built by humans, with zero after-sale support — now breaking. Over the past eighteen months, a handful of companies have demonstrated that satellites can be docked with, refueled, tugged to new orbits, and even repaired, by other spacecraft designed specifically for the job. The U.S. Space Force has committed over $140 million across four contracts to turn this capability from a demo into a procurement program.

Three companies won Space Force contracts in 2024–2026. Starfish Space collected a $37.5 million STRATFI for the Otter tug demo, a $52.5 million SDA disposal contract, and a $54.5 million APFIT for a dedicated GEO servicing vehicle — a combined $144.5 million in firm government commitments. Astroscale received $25.5 million for the Provisioner refueling spacecraft. Northrop Grumman secured two separate experiments through its SpaceLogistics subsidiary, plus the DARPA RSGS program launching this summer.

The case for commercial servicing

Starfish Space argues that satellite servicing is a commercial logistics problem, not a military one. Founded in 2019 by two former Blue Origin engineers, the company raised over $100 million in Series B funding in April 2026 — led by Point72 Ventures — and now counts NASA, the Space Force, the Space Development Agency, and Intelsat as customers. Its Otter spacecraft, at roughly one-tenth the size and cost of competing systems, uses electric propulsion and proprietary computer vision to autonomously approach and dock with unprepared satellites.

The Otter Pup 2 mission, launched in June 2025, is currently maneuvering toward a docking attempt with Gilmour Space's ElaraSat satellite in low Earth orbit — what would be the first-ever commercial satellite docking in LEO. If successful, it validates the core thesis: that servicing vehicles can approach and mate with unprepared commercial spacecraft using software alone, without expensive pre-installed docking hardware.

The commercial argument rests on unit economics. A geostationary communications satellite costs $200–$500 million to build and launch. A life extension mission from an Otter — which docks and provides station-keeping propulsion — costs a fraction of that. Intelsat, the first commercial customer, signed a contract in 2024 for exactly this use case. The logic is straightforward: why replace a functioning satellite when you can refuel it?

The case for military-led capability

The Space Force sees on-orbit servicing through a different lens: as a strategic imperative rather than a cost-saving measure. Satellite maneuverability is now a core requirement — the service's next-generation space domain awareness constellation, RG-XX, mandates on-orbit refueling by design. "I mandated that and was very militant about it," Maj. Gen. Stephen Purdy said in 2025. U.S. Space Command has called for all Department of Defense spacecraft to be equipped with refueling ports by 2030.

The military argument is about survivability, not cost. A satellite that can maneuver without regret, refuel on orbit, and be repositioned to dodge threats or cover new mission areas is substantially more resilient than one that must conserve every gram of propellant. Col. Scott Carstetter, director of SSC's servicing, mobility, and logistics office, told reporters in May 2026 that his office has the funding needed to transition the USSF-23 demonstrations directly into operational procurement.

"We're looking to operationalize these demonstrations to see how we can actually use them with operational clients that we have in need today," Carstetter said. The $54.5 million APFIT award to Starfish — the only space company funded in this APFIT cycle, and one of the largest APFIT awards in the program's history — signals that the Pentagon is already moving in that direction.

The infrastructure question

The deeper debate, and the one that will shape the industry's trajectory for the next decade, is about architecture. Refueling requires a fuel depot in orbit, a servicing vehicle that can dock with it, client satellites equipped with compatible ports, and — crucially — common interface standards. Without those standards, every refueling mission becomes a bespoke integration project, and the business case collapses.

Orbit Fab has shipped its RAFTI refueling ports to the Space Force and is building the fuel depot for the Tetra-5 demonstration on an Impulse Space hosting platform under a $13.3 million DIU contract. Astroscale's Provisioner will serve as the refueling vehicle. Quantum Space won a Department of War OECIF contract in June 2026 to build a fuel depot on its Ranger platform, with former NASA administrator Jim Bridenstine as CEO. Northrop Grumman's Passive Refueling Module was approved for inclusion on all DoD satellites.

The question is whether this evolves into an open, multi-vendor market — the model the Space Force says it prefers — or whether the combination of security requirements and interface lock-in concentrates the market around a single prime, as happened with launch vehicles before the commercial small-launch era. Starfish Space's strategy of building a small, software-defined vehicle that can dock with unprepared satellites is a bet on the open model: if the docking interface is universal, any servicing provider can compete.

The contrasting approach is Northrop Grumman's Mission Robotic Vehicle (MRV), developed under DARPA's RSGS program, which is a larger, more capable platform designed for heavier repair and assembly tasks rather than pure life extension. MRV can perform what the industry calls "complex servicing" — replacing electronics, repairing solar arrays, and assembling structures in orbit — but it requires satellites to be designed with servicing interfaces from the start.