The capability frontier in orbit is moving from one spacecraft acting alone to several acting together. Rendezvous and proximity operations — approaching and working close to another object in orbit — have historically been the job of a single, expensive servicer. The newer idea is to distribute the task across multiple cooperating spacecraft, and that is what this week's grant claims.
iMetalX's US12600498B1, "Multi-agent spacecraft system for rendezvous and proximity operations" (issued April 14, 2026), is tightly classified: B64G 1/242 (orbit/attitude control for rendezvous) and B64G 1/1085 (constellation/coordination). "Multi-agent" is the operative term — borrowed from robotics, it means several autonomous units coordinating toward a shared objective rather than one unit doing everything.
Here's the mechanism and why it's hard. Proximity operations are unforgiving: two objects moving at orbital velocity, closing to within meters, where a small relative-velocity error becomes a collision. Doing this with one spacecraft is already difficult. Coordinating several — so they don't interfere with each other, share a consistent picture of the target, and divide the work — multiplies the control and sensing problem. The payoff is capability: several smaller, cheaper agents can inspect a tumbling object from multiple angles, or jointly manipulate a structure too large for one servicer.
Why file on the multi-agent version specifically? Because the missions that justify on-orbit servicing — inspecting a failed satellite, refueling a fleet, assembling a large antenna or telescope in space — increasingly exceed what a single servicer can do, and because cheaper small spacecraft make a cooperating swarm economically plausible where one exquisite vehicle is not. The IP value sits in the coordination method, not the individual spacecraft.
The standard caution applies with extra force here. RPO is among the least forgiving, least flight-proven domains in spaceflight; a granted method describing multi-agent coordination is a long way from a demonstrated mission. The claim tells you how iMetalX proposes the agents cooperate — it does not tell you that a fleet has flown this, or how reliably the coordination holds when the target is uncooperative or tumbling.
For the servicing lane, the signal is the direction: away from the single exquisite servicer, toward cooperating fleets, with the defensible IP migrating to the coordination layer. Watch which assignees file multi-agent RPO claims — proximity operations are the gate to every in-orbit servicing and assembly business, and whoever owns the coordination owns a piece of all of them.