Company
Twenty-five years spent making spatial software reliable enough to ship in someone else's product. Three times now, the same problem has arrived in a new form — and each time, the answer has been a platform rather than a one-off.
Track record
Our founding team pioneered turn-by-turn navigation on consumer phones — years before it became a default feature. The company went on to list on NASDAQ in 2010, reaching roughly 100 million drivers.
Navigation delivered in the car itself, at automotive scale and to automotive standards, across roughly 50 million connected vehicles. This is where the discipline came from: shipping software that has to work for years in a product you do not control.
Five years, twenty engineers and roughly $22M to put indoor and outdoor autonomous parking into production. The hardest version of the problem, solved end to end — and the foundation of the platform.
The same core ported to smart shopping carts: a new embodiment, cheaper sensors and a dense human environment. Proof that the platform travels.
Navigation is the entry point, not the ceiling. As deployments accumulate, the platform becomes the shared spatial capability that robots across many categories are built on.
Where we sit
The Physical AI stack is settling into layers. Compute sits underneath. Foundation models push general capability forward, mostly around manipulation and task reasoning. Applications sit on top, in specific industries.
We work on the layer in between: spatial intelligence — knowing where you are, what is around you, where to go and how to get there safely. It is the capability every mobile robot needs and no robot company wants to spend three years building.
We build it as an open platform, available to any robot maker, rather than as part of a closed ecosystem around our own hardware. We do not build robots. That is deliberate: a platform only compounds if the people using it are not competing with its owner.
Licensed to any OEM, integrator or supplier — not reserved for our own products.
One core designed to travel across robot types, not one stack per robot.
Optimised for robots that must ship, run for years and make commercial sense.
Leadership
A team that has scaled navigation across roughly 100 million phones and 50 million connected vehicles, and has spent the last five years applying it to robots.
Localization and SLAM, perception, planning and control, deployment engineering. If you want your work running on real robots in real environments rather than in a demo, get in touch.