ORION ARM
The lunar surface from orbit, with Earth above the horizon

Where no one can help, the machine decides.

Autonomous robots for the lunar surface, and the on-board compute that lets them work without a lifeline.

Est. 2026 · China

Lunar surface — visualisation

01 — The conditions01 / 06

No crew.
No comms.
No resupply.

On the Moon those three are not edge cases. They are the baseline. A rover that waits for Earth is a rover that has stopped.

What is needed is a machine that perceives, decides and coordinates on its own — and a link that is allowed to be absent without the mission ending. Everything below follows from that single constraint.

This is the only problem we work on.

02 — The machine02 / 06

CHERI is not a probe.
It is a computing centre
that walks.

CHERI, a six-legged gold-foil rover
CHERI — engineering model, 5 kg
Mass
5 kg validated unit
On-board compute
20 TOPS
Autonomy
Offline perception, navigation, task planning
Comms
Store-and-forward; no live link required
Growth path
20 kg → 50 kg → 100 kg class

Images and science data are understood on the surface. Only conclusions travel home. Several units form a network and share position, task and compute — the nerve of a colony.

Two CHERI rovers on simulated lunar regolith, in front of a lava-tube entrance
Two units, simulated regolith — lava-tube entrance

03 — The plan03 / 06

Now

  1. L1 — Now

    Extreme-environment hardware

    Modules, single units and algorithms, sold outside deep space. Short cycles, real revenue, and it pays for the machines.

  2. L2 — 2027

    Complete robots

    Lunar rovers and surface construction vehicles delivered as whole systems rather than components.

  3. L3 — Mid-2030s

    Resource development

    The first closed loop: lunar material used where it is found, rather than shipped from Earth.

  4. L4 — Late 2030s

    Surface infrastructure

    Bases built and serviced by the same fleet — from a single mission to a long-duration presence.

  5. L5 — 2040+

    The colony

    Hundreds of nodes, self-organising across the surface. The largest machine colony off Earth.

04 — Capability04 / 06

Three things have to be true
at the same time.

01

Autonomous intelligence

Perception, navigation and task planning that run offline, on the same board as the sensors. No round trip to a ground station.

02

Flight-proven engineering

Twenty years of end-to-end spaceflight: design, integration, test, launch, operations. A whole machine is run, not assembled.

03

Extreme-environment hardening

Vacuum, thermal cycling, radiation, launch loads. The part that decides whether it survives the landing at all.

A concept of a lunar base with power masts, habitats and working vehicles
Surface infrastructure — concept

05 — Record05 / 06

Two national flagships.
One private company.

ORION ARM is the only private company selected for both Chang'e-8 and Tianwen-3 — China's two deep-space flagship missions — with early projects through ILRS, the International Lunar Research Station.

  • National programmes

    Delivery credit built on flagship missions, and the engineering process that comes with them.

  • Universities

    Zhejiang, Shenzhen and Beihang — instruments and custom builds for deep-space research.

  • Overseas

    Small, unit-level space nodes. High frequency, low mass, a delivery radius that reaches further than the domestic market.

06 — Contact06 / 06

Put the next machine
where you need it.

  • Business

    Payloads, complete vehicles, module supply, joint laboratories.

  • Press

    Technical briefings, imagery, cleared statements.

  • Careers

    Robotics, embedded systems, AI, structures and thermal, space systems.

Write to us

hello@orionarm.com