Sanctuary AI has announced a physical AI milestone with a global Tier 1 automotive supplier, demonstrating its technology on a wire-plugging task measured against the customer’s production benchmarks.
According to Sanctuary AI, the proof of concept achieved a task success rate above 99.5% and a cycle time of 2.54 seconds. The task involved manipulating flexible material and inserting a plug into a moving target on a conveyor. The company says the system matched the throughput of the customer’s live production line. Those results apply to this specific proof of concept and task; they should be understood in that context.
The work reflects an evolution in Sanctuary AI’s strategy: applying its physical AI to existing commercial robotic systems while continuing to develop capabilities for future hardware, including humanoids. This approach can give industrial customers a way to explore advanced automation using robotic platforms already available today.
Flexible objects and moving targets illustrate why dexterity matters. An industrial robot may need to respond to variations in position and movement while meeting demanding requirements for speed and repeatability. Demonstrating a system against a customer’s own benchmarks is an important step in understanding how such capabilities might fit into real operations.
We’re encouraged by Sanctuary AI’s focus on practical industrial challenges and look forward to following its continued development and customer work.
Sources & further reading
More updates
Sanctuary AI Appoints Daniel Friedmann as CEOFor educational purposes. This material is not investment advice or an offer to invest.

