Stereolabs Launches ZED Box 2 for Robot AI
The new hardware consolidates multi-sensor processing to enable real-time decision-making on autonomous machines.
Updated on Oct. 27, 2026 in Robotics
Live Poll
Would you trust a robot to make decisions using on-board AI rather than a data center?
Stereolabs has launched the ZED Box 2 Series, a new line of onboard AI computers designed to process multi-sensor data directly on robotic systems. The hardware integrates inputs from cameras, lidar, and motion sensors to enable real-time machine decision-making.
Why it matters
As robot fleets grow, builders face challenges in scaling high-volume sensor data processing while managing complex, multi-supplier hardware stacks. This platform aims to consolidate processing and sensor control into a single unified unit.
The ZED Box 2 standard model runs on an Nvidia Jetson Orin NX chip, while the Max model utilizes an Nvidia Jetson Thor chip. The series supports eight synchronized cameras and operates across a thermal range of -20 to 50 degrees Celsius.
The players
Stereolabs
An Ouster subsidiary specializing in visual intelligence and 3D sensing hardware for robotics.
Ouster
A provider of lidar sensing technology and digital lidar sensors for autonomous machines and robotics.
The details
The ZED Box 2 series functions as an edge computing node that aggregates data from cameras, lidar, and motor controllers on the machine itself. By performing AI inference locally rather than via cloud offloading, the hardware reduces data latency to allow for near-instant reaction times. The system is housed in a compact chassis measuring 55 millimeters in height to accommodate varied robotic form factors.
Timeline
October 27, 2026: Stereolabs launched the ZED Box 2 Series and Ouster shares surged.
The Tech Race
The integration of Nvidia Orin NX and Thor architectures highlights the intensifying race to place desktop-class AI capabilities directly onto mobile machines. This move positions Stereolabs to compete with custom-built sensor fusion boards by offering a pre-integrated alternative.
Robot developers can now adopt a unified computing package that removes the need for sourcing disparate motor controllers and sensor interfaces from multiple vendors. The system is available for integration into new and existing robotic fleets immediately.
The takeaway
The success of the ZED Box 2 will depend on how effectively it simplifies the integration of third-party sensor arrays for mid-sized robotics firms. Developers should monitor upcoming technical documentation to see if the hardware supports their specific lidar and camera communication protocols.
Further reading
For broader trends in autonomous machine hardware, see our Robotics coverage.
Source note: This article includes information reported by Benzinga.
Live Poll
Would you trust a robot to make decisions using on-board AI rather than a data center?
