TL;DR: The hardest space-debris problem is not only finding an object. Operators need observations they can turn into a timely, defensible decision. Across 1,382 measured answers, LeoLabs led the tracked providers at 10.6% answer presence. The shortlist changed sharply between sensing, collision decisions and active debris removal.
Want the same buyer-question baseline for your market? Start with the Schmitdy AI search audit.
Small debris can carry large consequences for a satellite operator.
The most difficult objects are too small for stable, routine tracking yet large enough to damage a spacecraft. A warning also has to arrive early enough to support a choice. The mission team must judge whether the object is real, whether its predicted path is reliable and whether a manoeuvre reduces more risk than it creates.
That makes the 1 to 10 cm problem a chain, not one sensor purchase. It joins detection, orbit estimation, conjunction assessment, operating rules, manoeuvre planning and, in a different class of mission, debris removal.
We tracked 55 questions from commercial satellite operators, constellation teams and European institutional buyers. Eight direct Project-S questions were kept outside the category ranking. The table below uses 1,382 answers to the remaining 47 questions across ChatGPT, Gemini and Google AI Overview.
| Tracked provider | Answers naming the provider | Share among tracked brand mentions |
|---|---|---|
| LeoLabs | 10.6% | 35.5% |
| Slingshot Aerospace | 5.4% | 22.0% |
| Neuraspace | 3.5% | 11.5% |
| ClearSpace | 3.1% | 9.4% |
| OKAPI:Orbits | 2.5% | 7.8% |
| COMSPOC | 2.2% | 8.4% |
| Astroscale | 1.5% | 5.8% |
| Project-S | 0.7% | 1.5% |
| Digantara | 0.1% | 0.2% |
Why is the 1 to 10 cm debris range so difficult?
Large tracked objects can receive catalogue identities and repeated observations. Very small particles may remain below the threshold for useful object-by-object decisions. The middle range is difficult because an object can pose a serious threat while observations remain sparse or uncertain.
Detection is only the start. A sensor must collect enough useful data to estimate the orbit and uncertainty. The observation must reach a processing system. The operator then needs a conjunction assessment that fits the spacecraft's orbit, manoeuvre ability, fuel limits and mission priorities.
The European Space Agency's space-debris work and annual Space Environment Report describe an orbital environment shaped by launches, fragmentation events, inactive spacecraft and spent stages. NASA's Orbital Debris Program Office publishes models, measurements and technical reference material.
Those sources help define the scale of the problem. They do not remove the operator's need to test one service against its own orbit, warning window and decision process.
Which providers entered the measured shortlist?
LeoLabs led this measured set at 10.6% answer presence and 35.5% share of voice among the tracked brands. Slingshot Aerospace followed at 5.4%. Neuraspace, ClearSpace, OKAPI:Orbits and COMSPOC formed the next group.
The table is not a claim that these companies sell the same product.
LeoLabs is associated with radar observations, tracking and space-domain awareness. Slingshot spans space-domain awareness and satellite operations. Neuraspace and OKAPI:Orbits appear around conjunction assessment, traffic coordination and manoeuvre decisions. COMSPOC combines tracking, analytics and space-domain awareness. ClearSpace and Astroscale are more closely tied to active debris removal and on-orbit services. Digantara develops space-surveillance infrastructure.
Project-S appeared in 0.7% of the category answers. Its strongest measured group was radar payloads and data at 1.7%. It appeared in 0.9% of debris-removal answers, 0.4% of SSA and STM procurement answers, 0.3% of debris-detection answers and none of the measured conjunction-assessment answers.
The engine split was also uneven. Project-S appeared in 1.1% of ChatGPT answers and 0.9% of Google AI Overview answers, but none of the measured Gemini answers.
The lesson for buyers is to compare providers inside the job they perform. The lesson for providers is to publish enough evidence for an answer engine to place them in the right part of the chain.
How should an operator compare sensing options?
Start with the orbit and the object range, not a generic accuracy claim.
Ground radar can revisit large parts of low Earth orbit and build repeated observations without placing another payload in space. Optical systems add useful coverage under different light and weather conditions. Space-based sensors can observe from another geometry and may reduce some ground-based limits, but they bring payload, downlink, calibration and mission-life constraints.
A serious test should ask:
- Which object sizes and orbits are in scope?
- How often can the system revisit the relevant volume?
- What is the delay from observation to usable state estimate?
- How does uncertainty change after each observation?
- How does the system behave after a fragmentation event?
- Which independent sources can confirm or challenge the track?
- Can the operator inspect provenance, timestamps and confidence?
Project-S publicly describes plans for high-sensitivity radar and algorithms aimed at the 1 to 10 cm range. Deutsche Welle's profile also describes a later robotic clean-up phase. These are company plans reported in public, not proof of live operating coverage or removal performance.
The right procurement step is a bounded demonstration. Use representative orbits, known objects and an agreed scoring method. Measure detection, track continuity, latency, false alerts and the effect on the final decision.
When does an observation become a collision decision?
A conjunction alert does not automatically mean a spacecraft should move.
The operator needs the predicted close approach, the uncertainty around both objects, the time remaining, spacecraft constraints and the cost of the manoeuvre. A burn can consume fuel, interrupt service and create another trajectory that must be screened. Waiting can leave too little time to act.
The decision process should state:
- which data sources enter screening
- how duplicate or conflicting tracks are handled
- what thresholds trigger review
- who owns the final decision
- how a candidate manoeuvre is checked against other objects
- how the team records the evidence and outcome
Automation can reduce repetitive screening and surface the cases that deserve attention. It should not hide uncertainty or make an unexplained recommendation. The operator needs to see which observation changed the risk and why the proposed action is safe.
This is where space traffic management differs from a simple tracking feed. The European Union Space Surveillance and Tracking partnership, EU SST, supplies collision-avoidance and related services to registered users. Commercial services can add sensors, analytics, interfaces and operating support. Buyers should test how well those parts fit the mission team's existing process.
What evidence should an SSA or STM provider publish?
Public evidence should connect a claim to a method.
A useful sensing page states orbit regimes, object ranges, coverage assumptions, update timing and validation. A conjunction page explains inputs, uncertainty handling, thresholds, human review and audit history. An integration page states formats, interfaces, security, retention and operating support.
Evidence can include:
- sensor specifications with test conditions
- calibration and validation methods
- research papers and conference material
- case studies with the orbit and decision window defined
- latency and availability measures with a time period
- independent comparisons or public mission results
- named team expertise tied to the relevant work
An early-stage provider may not have a full operational record. It can still publish a clear architecture, test method, current readiness level and limits. A precise statement of what has been demonstrated is more useful than a broad claim that sensing or artificial intelligence will solve debris.
Retrievability is part of that evidence. If a core page meets a browser challenge or blocks major crawlers, public answer systems may rely on posts, interviews and third-party summaries instead of the official description. The company then loses control over basic facts such as scope, readiness and buyer fit.
When does active debris removal enter the decision?
Active debris removal is not a substitute for every collision-avoidance manoeuvre.
Removal targets need technical, legal and economic assessment. A mission must identify and approach the object, understand its motion and condition, capture or stabilise it without creating more debris, and move it to a safe disposal path. Ownership, consent, liability and mission authority also matter.
Astroscale and ClearSpace appear in answers about removal and on-orbit servicing because they have built public category associations around those missions. A new provider needs separate evidence for sensing, autonomous operations, rendezvous, capture and disposal. Success in one part should not be treated as proof across the whole mission.
For most satellite operators, the near-term buying question remains mission assurance: better awareness, fewer false alerts, a shorter path from warning to decision and clear evidence after the event. Debris removal sits in a related but distinct institutional and service market.
What did the 1,382-answer source set trust?
ESA received 1,944 citations in the measured category set and NASA received 1,799. European Union sources received 396, ScienceDirect 394 and YouTube 259. Other cited domains included Orbital Radar, ResearchGate, MDPI, EU SST, the German Aerospace Center and Neuraspace.
The source mix makes sense. Buyers and answer engines look for public agencies, research, technical methods and mission evidence when the cost of a wrong claim is high.
It also shows the route for a commercial provider. Official source pages need to be accessible, specific and tied to evidence. Independent research and coverage should then test the claim. Company posts can distribute the result, but they should not remain the best public explanation of the product.
The 2026 decision
No single provider should win a shortlist merely because it appears most often across every space-safety question. The buying set should follow the mission job.
For sensing, test coverage, object range, track quality and latency. For collision decisions, test uncertainty, false alerts, workflow fit and manoeuvre support. For traffic management, test coordination, interoperability and audit history. For active debris removal, test mission readiness, capture evidence, authority and risk.
The 1 to 10 cm blind spot is a strong research problem because it exposes every weak link between observation and action. The provider that earns trust will show how its part of that chain works, where it stops and what proof supports the handoff.
Sources
- ESA Space Debris
- ESA Space Environment Report
- NASA Orbital Debris Program Office
- EU Space Surveillance and Tracking
- German Aerospace Center
- Deutsche Welle: Project-S profile
- LeoLabs
- Slingshot Aerospace
- Neuraspace
- Astroscale
- ClearSpace
If you want to measure the questions buyers ask in your category, book a 20-minute call with Marco: https://calendly.com/marco-ai-heroes/20min.
