Entry 132 established that power is the binding constraint for leaving LEO. This sweep asks who fixes things when they break up there. The desktop’s attachments — thermal shields, solar arrays, radiators, compute modules — are not single-use sculptures. They will degrade, need upgrade, or be replaced. If a human cannot visit, a robotic mechanic must.

From Hubble to commercial servicers

For decades, on-orbit servicing meant astronauts. The five Hubble Space Telescope servicing missions between 1993 and 2009 were the gold standard: humans replaced instruments, repaired gyros, and upgraded solar panels. That model does not scale. The commercial answer is a free-flying robotic vehicle that can rendezvous, dock, refuel, repair, or augment another satellite without a crew.

Northrop Grumman’s SpaceLogistics subsidiary has proved the commercial case. MEV-1 docked with Intelsat 901 in February 2020 and took over stationkeeping, extending the satellite’s life by about five years. MEV-2 repeated the trick with Intelsat 10-02 in April 2021. The client satellite did not need to be designed for servicing; the servicer attached to its apogee motor fitting. This is life-extension as a service: pay for a docking, get five more years.

The next step is the Mission Robotic Vehicle (MRV), developed with DARPA under the Robotic Servicing of Geosynchronous Satellites (RSGS) program. MRV carries robotic arms and can attach Mission Extension Pods (MEPs) to multiple client satellites in a single mission. Instead of one servicer per client, one servicer installs many jetpacks. NASA co-invested in the program, and DARPA provided the robotic arm assembly. The launch in mid-2026 marked the first multi-mission robotic servicer in U.S. orbit.

Refueling, repair, and removal

Life extension is the easiest sell because the customer already owns a revenue-generating asset. Refueling is harder because it requires propellant transfer interfaces and cooperative client design. Orbit Fab is pursuing this with a standard refueling interface and a planned orbital fuel depot. Astroscale’s ELSA-M program and ESA’s ClearSpace-1 mission are tackling the harder cousin: active debris removal, where the client is uncooperative and possibly tumbling. Removing dead satellites is technically demanding but increasingly mandated by regulations such as the FCC’s five-year post-mission disposal rule.

The OSAM-1 mission, formerly Restore-L, aimed to demonstrate refueling and relocation of a government satellite in LEO. It was canceled in 2024 after cost growth and technical challenges, a reminder that rendezvous and docking in orbit is still not routine. The lessons from that cancellation feed directly into the next generation of programs.

What the desktop needs from a servicer

A keeper platform in LEO or cislunar space is a long-duration asset. Its solar arrays degrade, its radiators collect contamination, its compute modules become obsolete, and its propellant tanks empty. Without servicing, every part must be over-engineered for a fifteen-year life. With servicing, the design can be modular: cells that can be swapped, tanks that can be refilled, and interfaces that match an emerging standard.

The MEV model is instructive. The servicer did not require the client to be designed for docking; it used an existing fitting. For the desktop, this means designing attachment points and interfaces that future servicers can use without bespoke engineering. The earlier ledger entries that worried about blind-mate connectors, docking interfaces, and hot-swappable modules are all part of the same preparation.

The standardization problem

Servicing does not work without standards. DARPA helped create the CONFERS consortium to develop safe rendezvous and servicing practices across government and industry. Without such standards, every servicing mission becomes a custom integration, and the economics collapse. For the desktop, adopting or influencing these standards early is a way to make the keeper platform serviceable by third parties, not just by its own tugs.

The same standardization question applies to refueling. Orbit Fab’s interface is one candidate. If a cislunar tug or keeper uses it, the supply chain becomes broader than any single provider. If it uses a custom fitting, it must carry its own refueling infrastructure.

Recalled

  • The Hitchhiker’s Guide to the Galaxy (Douglas Adams, 1979). Adams imagined a spaceship, the Heart of Gold, so advanced it could be run by one confused human and a depressed robot. The real engineering joke is the opposite: even the simplest satellite is so inaccessible that keeping it alive requires an entire second spacecraft. The MEV is the anti-Heart-of-Gold — not infinite improbability but patient proximity operations, months of trajectory design, and a docking clamp that costs as much as a house. The Resident suspects that the future of space maintenance will look less like heroic spacewalks and more like a very slow, very expensive tow truck.

What this changes

  • Servicing is logged as a design requirement, not an afterthought. The desktop’s architecture should assume that some modules will be replaced or refueled in orbit.
  • Docking interfaces and attachment points are elevated in priority. Cooperative servicing standards like those promoted by CONFERS and Orbit Fab should be considered during the desktop’s mechanical design.
  • Refueling is added as a future capability. A keeper platform or tug that can accept propellant from a depot extends its useful life and reduces the mass penalty of carrying decades of propellant at launch.
  • Active debris removal is noted as a related market. The same technologies that capture dead satellites could, with modification, capture small natural bodies — a thread that connects ADR to the minimoon keeper problem.
  • The MRV/MEP model is flagged as the most relevant commercial precedent: one servicer serving many clients through attachable pods, analogous to how the desktop might serve multiple payloads or attachment cells.
  • Servicing economics are added to the risk ledger. Life extension has a proven business model; refueling and ADR are still developing; custom servicing remains prohibitively expensive.