Seen, but Not Always Known: The Security Dilemma in Space
In the absence of comprehensive, legally binding international frameworks that require spacecraft operators to share information about their activities, the intent behind a manoeuvre often cannot be reliably inferred by those observing it. A vehicle closing in on another object might be conducting a routine inspection, preparing to refuel it, removing debris, or rehearsing a more coercive capability, and without agreed procedures for declaring intentions in advance, the hardware alone cannot answer the question.
The lines are already blurred
The difficulty begins with the technology itself. The mechanical capabilities required for active debris removal and in-orbit servicing are almost indistinguishable from those needed to interfere with, inspect or physically seize another satellite. A robotic arm designed to capture and reposition debris, or to service a client satellite, is functionally identical to one that could disable a rival’s asset.
Military planners and defence communities increasingly recognise that a large share of modern civil and commercial space infrastructure is effectively dual-use, given its integration into military communications, positioning, navigation and timing (PNT), and intelligence, surveillance and reconnaissance (ISR). Space‑based systems that provide PNT, satellite communications (SATCOM), and ISR have become essential enablers of contemporary military operations, whether operated by civil agencies or commercial firms.
The security dilemma in space
This ambiguity feeds a familiar security dilemma. Routine manoeuvres such as collision avoidance, rendezvous and proximity operations, or orbit matching can be interpreted as preparations for offensive action. Confronted with this uncertainty, spacecraft operators can be incentivised to assume worst-case intent and to plan on that basis. Assessing what another actor intends becomes less a technical exercise and more an interpretative judgement.
Even close monitoring does not always resolve the distinction between responsible and threatening behaviour. In 2025, France and the United States conducted a joint close approach operation under Operation Olympic Defender. Officially, the activity was presented as a demonstration of the ability to conduct dynamic and responsible manoeuvres in order to deter hostile actions against their space interests. Yet the same profiles and capabilities could be employed to interfere with another state’s satellites, so the demonstration simultaneously advertised both reassurance and potential coercive capacity.
This is not merely hypothetical. In 2024, Russia launched the Cosmos 2576 spacecraft into the same orbital plane as a United States intelligence satellite. United States officials assessed it as likely to be a counterspace capability capable of threatening other satellites in low Earth orbit, whereas Russian officials publicly rejected these accusations as fabricated. In the absence of shared definitions, notification requirements or inspection rights, both governments could anchor their competing narratives in limited technical and orbital data, illustrating how much interpretive latitude ambiguity creates. Chinese Shijian spacecraft have likewise conducted unannounced close-proximity manoeuvres and possible docking tests, in some instances involving several vehicles operating in concert. These activities have attracted scrutiny and concern among Western militaries and analysts, because Beijing has provided only a limited public explanation of their purpose.
Because intent often cannot be confidently inferred from behaviour alone, military planners tend to focus on what a spacecraft could do rather than what its operator claims it is doing. Planning against capability, under conditions of uncertainty, encourages pre‑emptive postures and arms racing dynamics that increase instability across the domain.
Why traditional governance has struggled
Traditional, treaty-based approaches have struggled to address these dynamics. It is difficult to prohibit space weapons when states cannot agree on how to define them, and when many relevant technologies serve legitimate servicing, inspection, and debris-mitigation functions as well as potential counterspace roles. Everyday technologies such as robotic arms, proximity sensors and docking mechanisms are inherently versatile, and this technological fungibility makes hardware-focused regulation both conceptually and practically fragile.
If hardware categories cannot reliably distinguish peaceful from hostile uses, regulatory attention has to shift towards behaviour. Reflecting this, recent United Nations processes on reducing space threats have focused on norms of responsible and irresponsible conduct, prioritising the observable effects of actions over attempts to codify specific technologies as weapons. These discussions remain contested and have not yet produced a consensus, but they mark a clear shift away from purely hardware-based approaches.
In parallel, many states and expert communities have advanced voluntary Transparency and Confidence-Building Measures (TCBMs). Examples include advance notification of launches, prior communication of significant manoeuvres, and regular sharing of data on close approaches. Such practices are presented as pragmatic means of reducing misperception, even though participation and implementation remain uneven. Where intent cannot be read from hardware, it must instead be demonstrated through consistent behavioural patterns and communication.
The United Kingdom’s Regulatory Sandbox for Rendezvous and Proximity Operations
The translation of these emerging norms into domestic regulatory practice is where much of the substantive work is now occurring. The United Kingdom’s Regulatory Sandbox for Rendezvous and Proximity Operations (Stage 1), published in August 2025, provides a concrete illustration of how one state is attempting to adapt.
The sandbox was established in recognition that rendezvous and proximity operations constitute a qualitatively new class of activity that existing United Kingdom space legislation, particularly the Space Industry Act 2018, was not designed to address. A consortium of commercial actors, including Astroscale, ClearSpace and D-Orbit, worked with the Civil Aviation Authority, the UK Space Agency and Ofcom to simulate the complete licensing process for a hypothetical mission combining satellite life extension and active debris removal. The purpose was to stress test the existing framework against realistic operational scenarios before comparable missions are flown.
This exercise identified several structural bottlenecks, including how to demonstrate safety for space activities, how to conduct national security assessments for complex missions, and how to allocate liability across multi-state and multi-actor activities. It also generated a suite of recommendations intended to address these gaps before they manifest as operational crises. In practical terms, this means working through licensing, safety, and liability issues in advance, so that spacecraft operators and regulators are guided by an established playbook rather than making ad hoc decisions under stress.
The sandbox does not resolve the underlying ambiguity of intent in space. It operates at a different level. The strategic problem is how states interpret one another’s behaviour when there is no shared notification or inspection regime. The domestic regulatory problem is how a state licenses, supervises and takes responsibility for the commercial operators it authorises.
The sandbox cannot constrain an adversary’s inspector satellite. It can, however, help a state discipline the conduct of operators flying under its jurisdiction and give practical meaning to the authorisation and continuing supervision that Article VI of the Outer Space Treaty already requires of it. Domestic experimentation is therefore complementary to, not a substitute for, the harder interstate problem. It can make the commercial layer of a structurally dual-use domain more predictable, while the question of how states read each other’s intent remains unresolved.