04 ยท The edge

Everyone else flies one bespoke servicer per target. We coordinate a swarm.

On-orbit servicing already exists, but every operator sends a single custom vehicle to dock with a single cooperative satellite. That can’t reach 20,000 targets, and it can’t touch objects never built to be grabbed. Our difference is structural: many small robots that coordinate as one to move any object, with a policy that gets cheaper and better every mission.

WhoApproach todayWhere it stops
Northrop GrummanMEV / MRVA tug docks to one satellite to extend its life.One vehicle per cooperative target, with no manipulation and no fleet.
AstroscaleJapan / UKSingle servicer for debris removal & life extension; ELSA-d demonstrated.Bespoke craft per job; no multi-robot coordination or shared learning.
ClearSpaceSwitzerlandESA-backed single-target debris capture (ClearSpace-1).One mission, one object, European-scoped.
MomentusUSAOrbital-transfer “space tugs” that ferry payloads to new orbits.Point-to-point delivery, not on-orbit manipulation; operational setbacks.
StelarisA swarm of small robots that coordinate to grab and reposition any object, cooperative or not.Software-first: cost per mission falls and capability compounds with every flight.

Swarm, not a ship

Many units jointly move a shared mass from only local information: the one capability no incumbent has. It works on objects never designed to be captured.

Software-first

The asset is the coordination policy, not a one-off spacecraft. Target $5–20M per mission against $30–60M for bespoke servicers.

It compounds

Every mission trains the fleet. The lead widens with operations: a coordination-and-data moat no competitor can buy. It only accrues by flying missions.

Orbit Fab (on-orbit refueling) is a complement, not a competitor. A swarm would refuel at its depots. The coordination kernel behind this is already validated in high-fidelity simulation: see the system →