Physical AI for space · Zürich

The autonomous workforce for the orbital industrial economy.

Everything big humanity builds in space over the next two decades gets assembled in orbit, not launched whole. Shipwright builds the machines that will do it.

SCROLL ▾
0
publicly documented, fully autonomous captures of an unprepared, uncooperative satellite — completed to date
€100M
the cost of today’s first-of-kind capture and servicing missions, built and verified largely as bespoke systems
3
recent commercial lunar landing attempts hit by altitude-sensing failures — two were lost; one landed hard and tipped
2030
ESA Zero Debris deadline — the capture and inspection missions are being funded now
The Gap

Launch cost is collapsing. Fairing diameter isn't.

The binding constraint in space is shifting from mass to structure size — and every operation that is limited by structural size (servicing, debris capture, module swap, assembly) ends with a close approach to an object that is unprepared, and silent.

Cooperative docking sensors have been off the shelf for twenty years. The autonomous, uncooperative version has never existed — piloted operations take months and ground crews, and autonomous attempts keep failing.

Shipwright is closing that gap is where the orbital economy starts. That is the work.

Step One · Helm

First, teach the machines to see.

Helm is a perception head and embedded autonomy stack for rendezvous and proximity operations against uncooperative targets. Perception is the capability everything else stands on — nothing can be captured, serviced, or built by a machine that cannot see. So that is where we begin.

Relative navigation

Range, pose, and motion of uncooperative targets — no beacons, no markers, no cross-links, no GPS. Built for the environments where all four are missing.

Target characterization

Geometry, tumble state, and berthing-feature identification — so you know what you're approaching before you commit the vehicle to it.

Assurance engineered in

Uncertainty bounds on every estimate, safety envelopes around every maneuver, and abort as a verifiable behavior.

Auditable by design

Every decision logged and signed. Evidence for your review board, your insurer, and the compliance rules the draft EU Space Act is bringing.

A component, not a project

Integrates on any host vehicle, inside any GNC loop. The assurance case transfers with it — no per-mission re-qualification, no bespoke re-engineering. Autonomy as a product, not a one-off.

Heritage, mission by mission

Helm rides partner missions as a hosted payload or redundant relative-nav sensor — earning flight heritage alongside proven chains, one flight at a time.

The Roadmap

Eyes, then hands, then the worker, then the shipyard.

The destination is orbital assembly — machines that build and maintain infrastructure in space. Nobody gets there in one leap. Each phase is proven in flight before the next depends on it: the eyes make the hands possible, the hands make the worker possible, and fleets of workers make the shipyard.

PHASE 01 / NOW

Helm · the eyes

Perception & GNC · the phase we're in now

Relative navigation and target characterization for GPS-denied, uncooperative environments — with assurance engineered in. Flying as a component on partner missions across servicing, debris removal, and inspection: every flight compounds the operational experience the next phase is built on.

PHASE 02 / NEXT

Grapple · the hands

Capture & manipulation · built on the Keel platform

Capture and module-swap manipulation for debris removal and orbital hardware refresh — the hands, guided by flight-proven eyes. Grapple is built on Keel, the common spacecraft platform every Shipwright machine shares from here on.

PHASE 03 / LATER

The Rigger · the worker

The complete unit · eyes + hands + Keel

Helm's eyes, Grapple's hands, and the Keel platform integrated into a single autonomous unit. A Rigger approaches, captures, and works on structures unsupervised — the complete orbital worker, every part of it already proven in flight.

PHASE 04 / DESTINATION

The Shipyard

Orbital assembly · fleets of Riggers

Robotic assembly of everything too big to launch whole: orbital data-center radiators and solar arrays beyond deployable limits, persistent manufacturing platforms, kilometre-scale apertures — built by fleets of Riggers working in concert.

Assurance

No one lets a machine near their spacecraft on trust alone.

A proximity operation puts our autonomy within metres of an asset worth hundreds of millions. Before any operator allows that, they need proof the machine will behave — so assurance is engineered in, not bolted on: every estimate ships with uncertainty bounds, every maneuver stays inside a safety envelope, and abort is a verifiable behavior, demonstrated under injected faults.

And every decision is logged and signed. When a review board, an insurer, or a regulator asks how the system behaved at closest approach, the evidence already exists.

Why Now

The window for this company is open now.

/01

Debris mandates are becoming missions

ESA's Zero Debris 2030 targets are converting into funded capture and inspection missions today — every one a proximity operation against an uncooperative target, every one needing the eyes.

/02

Autonomy is the unsolved layer

No one has ever completed an autonomous docking with an unprepared satellite. Piloted operations work but take months and ground crews. The delta between those two rows is where the next two decades of orbital work get unlocked.

/03

Regulation will demand verifiable autonomy

The draft EU Space Act brings orbital traffic rules that require compliance evidence — autonomous behavior that can be audited, bounded, and trusted. That standard is exactly what we build to.

/04

Cheaper missions mean more missions

Falling launch costs multiply mission cadence — and every additional mission needs the perception-and-autonomy layer that rides along. The economics of the vision improve with every launch-price drop.

Why Us · Why Switzerland

Built in Zürich, at the heart of European robotics and autonomy.

Founders with autonomous-systems, agent-trust, and quantitative-risk backgrounds — the disciplines this problem is made of. Switzerland is at the heart of space and robotics.

Europe is funding orbital servicing, debris removal, and sovereign space infrastructure at a scale it never has before — and it needs its own supplier for the autonomy those missions run on. That is the company we are building.

Get In Touch

Flying a proximity mission?
We should talk.

We're working with servicers, ESA programs, and platform builders on first flights — and building toward Europe's first low-cost uncooperative capture demonstration. If your mission has to see, approach, or touch something in orbit, we should be talking.

info@shipwright.space

Shipwright · Zürich · Switzerland