Earth orbit has long been a sparsely populated territory, but this shared ecosystem is increasingly under stress.
SpaceX now operates more satellites than the rest of the world combined, and in January announced plans to begin offering Stargaze, a free conjunction screening service built on its own on-orbit sensing network. At the same time, the U.S. Office of Space Commerce is proceeding with development of TraCCS, a vendor-neutral traffic coordination system, and Chinese space leaders are publicly warning about megaconstellation risk while accelerating their own deployments.
Safety services are allegedly expanding in parallel with congestion — I say “allegedly” because so far, there’s more talk than substance. But these plans also reveal a structural risk we have not fully addressed: The space community is drifting toward a coordination model that depends on a small number of dominant sensing and screening layers, which is, any way you slice it, biased.
When an event like a traffic accident or violation occurs, arriving at what’s true is based upon a number of eyewitnesses and evidence. Each offers one interpretation of what happened. Diversity and independence of observers yields the most accurate and precise perspective, a holistic one where consensus and mutual agreement can surface and divergence can be explored.
But if the number of independent and diverse observers decreases, the amount of unchecked biases rises and becomes dominant. It’s not just the number of observers that matters, but the diversity of perspectives. If we have many observers, but they’re all the same — the same values, same models, same type of sensor — then systematic biases go without verification. Truth is what surfaces from tension and scrutiny. We should require that for our own safety.
A more resilient approach would be to make disagreement among independent tracking systems visible and actionable, instead of quietly collapsing it into a single fused picture. Orbital safety should be built on structured divergence across many diverse actors, not soft dependence on any one of them. A Lord of the Rings approach — one space situational awareness system to rule them all — would be irresponsible.
Large constellations generate not only traffic but data. Operators with thousands of spacecraft have dense telemetry streams, frequent orbit updates and onboard sensing. When an operator provides conjunction screening to others at no cost, the benefits are obvious immediately: Smaller satellite owners gain access to higher cadence alerts and better situational awareness than they could build alone. Entry barriers drop. Coverage improves. So, this part is good — “a rising tide lifts all ships” sort of thing.
The danger is that dependency risk grows concurrently. When many operators rely on one operator’s screening outputs, that entity’s models, assumptions, and blind spots propagate through the ecosystem, unchallenged by the scrutiny that naturally occurs as the result of having numerous independent and diverse opinions.
My point isn’t that Stargaze is bad or that SpaceX is acting in bad faith — quite the opposite, in fact. Rather, the risk I see is in killing off plurality and independence of evidence if we sit back and let SpaceX become the sole, or dominant, provider of space situational awareness services. When safety becomes shaped by one dominant epistemic layer, that is fragile by design. Free screening improves access, yet it also concentrates coordination influence in ways the current governance model does not offset. Who’s going to check SpaceX’s answers or opinions? Many safety-critical fields already guard against this kind of fragility by design. Aviation, cybersecurity, finance, and weather forecasting all rely on independent verification and multimodel or multisensor cross checks, not because the leading provider is suspect, but because resilience requires plural, competing lines of evidence rather than a single dominant epistemic source.
Most of the current and proposed space traffic coordination concepts assume that safety improves as more raw data are pooled and fused. In practice, many operators can’t share full measurement data, covariance details or processing pipelines. National security limits, proprietary algorithms and liability exposure all restrict disclosure. Fusion works best in high-trust environments, and orbit is not one.
I propose a different model. Instead of pooling measurements, each operator would publish bounded uncertainty envelopes around its own satellite orbital estimates. These envelopes express what remains plausible and what is excluded, without exposing raw inputs. Coordination is triggered when independent envelopes diverge beyond safe limits for a conjunction assessment. Here’s an example: Operator X has one space situational assessment of a conjunction, Operator Y another. Each can provide plausibility bounds on its assessments. That is to say, each can state that a given conjunction can occur within some window of time and place, with no mention of likelihood or probability and with no attribution of responsibility or correctness.
The operators would then reconcile differences through a structured envelope-comparison protocol: When plausibility bounds overlap within agreed safety margins, coordination proceeds conservatively; when they conflict, a predefined escalation path triggers targeted follow-up observations, model reruns, or third-party arbitration to shrink the envelopes until operational agreement is reached. This keeps proprietary data private while still enabling cross-checking, convergence, and accountable resolution through independent evidence streams.
Technically, this is lighter weight than full data fusion and more compatible with mixed civil, commercial and defense participation. Architecturally, it avoids a single clearinghouse of truth. Coordination can be driven by shared uncertainty bounds and visible divergence, without requiring universal raw data disclosure.
Today, disagreement between tracking systems is often hidden inside fused catalogs and averaged solutions. Operators see a recommended miss distance and a confidence metric and never see the spread of underlying model outputs. When estimates suddenly shift or alerts arrive late, the prior divergence becomes invisible. Making divergence explicit changes behavior. If independent tracking pipelines disagree beyond defined thresholds, that signal surfaces early. Analysts investigate sooner, and operators coordinate sooner. Model weaknesses are exposed before they become operational surprises.
Orbital congestion is now a present condition, not a future scenario. Safety will depend less on whether we can produce one perfect picture and more on whether we can recognize when our pictures don’t match. Independent sensing, independent estimation and shared uncertainty bounds allow coordination without forced transparency and without concentrated epistemic control. In a crowded orbit, many eyes with visible differences form a stronger shield than one dominant “eye of Sauron” view.
