Launch is layer zero
An economy begins when assets can change owners, locations, configurations, and lifetimes through repeatable transactions. Cheap insertion only places inventory on the shelf.
Launch and insertion
Affordable access to every inclination and high-energy orbit
- Current boundary
- Routine: High-cadence rideshare and dedicated launches
- System metric
- $/kg to the required orbit
Orbital transfer
Transparent price, schedule, and delivered-mass data across orbits
- Current boundary
- Early service: D-Orbit reported 21 commercial ION missions and more than 200 payloads delivered by November 2025
- System metric
- $/kg per delivered Δv
Relative navigation
Routine autonomous capture across unknown target states
- Current boundary
- Demonstrated: ADRAS-J inspected an uncooperative rocket body and approached to about 15 m
- System metric
- Successful approaches without intervention
Tracking and traffic data
Persistent custody of small objects and shared manoeuvre protocols
- Current boundary
- Routine: Public and commercial catalogues support conjunction screening
- System metric
- Detection threshold and warning quality
Power and storage
Utility-like shared power, connectors, billing, and redundancy
- Current boundary
- Routine: Solar arrays and batteries support individual spacecraft
- System metric
- $/delivered kWh-year
Communications
Comparable delivered cost per GB for orbital customers
- Current boundary
- Routine: Large commercial networks sell continuous connectivity
- System metric
- $/GB at stated latency and availability
Rendezvous and docking
Broadly adopted robotic interfaces outside crewed vehicles
- Current boundary
- Early service: Two MEVs docked with commercial GEO clients; MEV-1 later undocked and moved to another client
- System metric
- Dockings per vehicle and anomaly rate
Fluid transfer
The planned on-orbit methane transfer did not occur; no recurring bulk service
- Current boundary
- Demonstrated: RRM3 stored cryogenic methane for four months without boil-off and exercised transfer tools
- System metric
- kg transferred, loss fraction, and $/kg
Robotic service and repair
Routine component replacement, repair, and upgrade
- Current boundary
- Early service: Commercial propulsion takeover has delivered more than ten combined client-years
- System metric
- Client-years added per intervention
Assembly and manufacturing
Published throughput, yield, and advantage after launch and return
- Current boundary
- Demonstrated: ISS manufacturing experiments and repeated commercial return capsules
- System metric
- Saleable output per machine-hour
Disposal and re-entry
Recurring removal of unprepared objects; ClearSpace-1 is planned for 2029
- Current boundary
- Planned: Close inspection, controlled re-entry, and cooperative deorbit are proven separately
- System metric
- $/kg removed and disposal success rate
The scarce product after cheap launch is not mass in orbit. It is a useful asset-year in the right orbit, with power, data, mobility, maintenance, and a credible way out.
Mobility before manufacturing
Rideshare lowers the price of reaching one drop-off orbit. A customer still needs last-mile deployment, phasing, plane changes, station-keeping, collision avoidance, and sometimes a return trajectory.
Delta-v is the orbital freight bill
The rocket equation makes velocity change multiplicative: propellant fraction rises exponentially with required delta-v and falls with exhaust velocity. Electric propulsion trades thrust and delivery time for propellant efficiency; chemical propulsion buys speed with mass. There is no universal “cost per kilogram in orbit” without a destination, deadline, and disposal requirement.
What repeat service looks like
These are provider-reported operational milestones. The missing market statistic is price per kilogram delivered to a defined orbit with a defined schedule and reliability.
Orbital power starts with area, not price
At Earth's distance from the Sun, a perpendicular surface receives about 1,361 W/m². Conversion efficiency, eclipse, pointing, degradation, and storage determine how much array area turns into dependable load.
56,940 kWh delivered in one year at the selected availability. This is an energy balance, not a spacecraft mass or cost estimate.
Servicing changes the spacecraft before it changes the mission
A client designed for capture, fluid transfer, modular replacement, and safe isolation is cheaper to service than a machine a robot must first interpret and disassemble.
The demonstration-to-service record
- 2020
MEV-1 docks with Intelsat 901
A commercial servicer takes over propulsion and attitude control for a client that was not designed around a new robotic interface.
- 2021
MEV-2 docks in operational GEO
The second mission attaches without first moving the client to a graveyard orbit, turning a first into the beginning of a product line.
- 2024
ADRAS-J closes on unprepared debris
A commercial spacecraft characterises and circles a roughly three-tonne spent upper stage, later approaching to about 15 metres without capturing it.
- 2024
OSAM-1 is cancelled
NASA ends a mission intended to refuel an unprepared Landsat after technical, cost, schedule, and market-alignment problems.
- 2024–25
Commercial return repeats
Varda returns W-1, W-2, and W-3 capsules; W-1 processed ritonavir crystals, proving return logistics rather than profitable factory throughput.
- 2026
MRV reaches orbit
Northrop Grumman launches a robotic servicer with three mission-extension pods. Launch is measured; repair and pod-installation performance remain to be demonstrated on clients.
OSAM-1 is evidence, not a footnote
The mission attempted autonomous rendezvous, robotic access to an unprepared fuel valve, refuelling, relocation, antenna assembly, and beam manufacture in one architecture. Cancellation does not show that the tasks are impossible. It shows that combining immature interfaces into one bespoke mission can erase the value of cheaper launch.
When does repair beat replacement?
Servicing wins when the expected value of added life exceeds the mission, integration, failure, and downtime burden. The model exposes the boundary; it does not estimate any operator's confidential contract.
Servicing has the higher expected value
- Expected preserved value
- $391m
- Servicing net value
- $312m
- Replacement net value
- $75m
- Failure exposure
- $34m
Calculation & assumptions
Servicing net value = annual value × added years × probability of success − service cost − annual value × downtime / 12. Replacement net value = annual value × the same comparison window − replacement and relaunch cost. A real decision also includes insurance, residual spacecraft life, schedule risk, financing, spectrum rights, and the probability that a new satellite earns more than the old one.
The missing standards layer
The International Docking System Standard defines a physical interface for crewed spacecraft and collaborative missions. It proves that compatibility can be designed. It does not supply a universal small-satellite grapple fixture, refuelling port, electrical connector, data protocol, or safe state for robotic repair.
Mechanical capture
Robots need known load paths, keep-out zones, fiducials, and a fixture that survives launch but remains accessible after years in space.
Fluids
Propellants differ in chemistry, pressure, temperature, phase, seals, and contamination tolerance. A standard port is only the visible end of a compatible tank system.
Power and data
A replaceable module needs connector geometry, voltage, communications, authentication, and fault isolation—not merely matching bolt holes.
Operations
Permission to approach, responsibility during joint control, cyber trust, collision liability, and evidence for insurers must be repeatable too.
The commercial breakthrough is therefore partly architectural: cooperative clients can make the servicing spacecraft simpler, smaller, and reusable across a fleet.
A factory needs yield, not just microgravity
Orbit can remove sedimentation and buoyancy-driven convection, expose materials to vacuum, and allow structures larger than a fairing. None of those advantages automatically pays for equipment, crew time, power, quality control, launch, and return.
| Claim | Measured evidence | Commercial unknown |
|---|---|---|
| Make a part in microgravity | Tools, polymers, ceramics, fibres, and biological samples have been produced on ISS | Repeatable yield and machine utilisation |
| Process a product autonomously | Varda's W-1 reported growing Form III ritonavir crystals | Value attributable to microgravity versus terrestrial processing |
| Return payloads | W-1, W-2, and W-3 completed re-entry by May 2025 | Return price, cadence, insurance, and recovered yield |
| Assemble beyond a fairing | Robotic and crewed assembly have extensive station heritage | Autonomous free-flyer throughput and customer demand |
The unit that matters is saleable output
Manufacturing throughput should be reported as accepted product per machine-hour, with scrap, crew intervention, power, consumables, and return included. “Manufactured in space” is a location; “manufactured profitably in space” is a system result.
Disposal is part of every sale
An orbital economy consumes a shared, finite environment. Every asset needs collision avoidance while alive, passivation when dead, and either prompt re-entry or a stable disposal orbit.
The environment is already an operating cost
The FCC requires covered satellites ending missions in or passing through LEO to dispose within five years. Compliance and controlled re-entry reduce new risk; neither removes the large unprepared objects already in orbit.
Inspection is not removal
ADRAS-J demonstrated autonomous proximity operations around an uncooperative upper stage and validated collision avoidance, then began its own deorbit. Capture and removal are a later mission. ESA's ClearSpace-1, now planned for 2029, is intended to remove the 95 kg Proba-1 satellite. The sequence is honest engineering: find, characterise, approach, capture, control, then dispose.
What a mature orbital supply chain would publish
Service economics
Price per kilogram delivered to a stated orbit and delta-v; price per kilogram of propellant transferred; client-years added; interventions per servicer; manufacturing yield and accepted output.
Reliability
Autonomous operating hours, mean time between human interventions, docking attempts per success, fluid loss, repair closure rate, disposal success.
Infrastructure quality
Delivered kWh-year, communications cost per delivered GB, navigation availability, conjunction-warning quality, compatible interfaces and suppliers.
Market depth
Repeat customers, competing providers, contracted cadence, utilisation, insurance terms, and revenue that does not depend on a demonstration subsidy.
The constraint after cheap launch
Customer density has to rise faster than complexity
A tug, depot, repair robot, or return capsule becomes infrastructure only when enough compatible customers share its development and standing capacity.
Last-mile mobility
Rideshare creates many payloads with nearby logistics needs and supports repeat tug missions.
GEO life extension
Large, revenue-producing clients make the value of several added years unusually clear.
Depots and factories
They need compatible fleets, high utilisation, transparent unit economics, and reliable return or disposal.
Evidence, boundaries, and sources
This report treats a flight as measured only after the relevant operation occurs. Provider mission counts are labelled as company-reported. Announced launch dates and intended capabilities are targets. Editorial maturity ratings compare repeatability and market depth; they are not official NASA technology-readiness levels.
- Measured
- A completed flight event, agency statistic, or physical quantity reported by a primary source.
- Company-reported
- An operational result published by the provider and not independently audited here.
- Derived
- A calculation shown with its inputs and boundary, such as the solar-area model.
- Target / planned
- A future mission, capability, price, date, or performance claim.
- Editorial inference
- A maturity classification or conclusion drawn across the evidence.
