Electronics distributor Avnet and Singapore robotics integrator Weston Robot said they are bringing a jointly developed edge-AI inspection platform to industrial customers, mounting an Avnet compute backpack on top of a Weston-integrated quadruped robot to patrol factories, energy assets and other complex facilities autonomously.
What the platform does
The combined system pairs Weston's fleet-management software with an Avnet compute payload capable of delivering up to 50 TOPS of AI performance on-device. That is enough headroom for the robot to run thermal and visual analytics, 3D lidar mapping and anomaly-detection models along a defined inspection route, and to keep operating in GPS-denied environments such as substations, chemical plants and underground infrastructure where cloud connectivity is unreliable.
AMD silicon inside the backpack
The Avnet reference architecture is built around AMD Ryzen AI Embedded processors, executing inference on the edge to cut round-trip latency and reduce reliance on backhaul. New inspection or perception tasks can be trained off-line and re-deployed to the robot, extending capabilities beyond what the base Unitree quadruped ships with.
Industry timing
The launch lands as quadruped inspection is moving from pilots to deployments, with Boston Dynamics' Spot on paid patrols and warehouse operators standardising on collaborative mobile robots. Weston's RaaS model — where operators pay for the fleet by subscription rather than up-front — mirrors the buying pattern DHL and other logistics giants have used to scale to thousands of robots without heavy capex.
Target industries
Avnet and Weston are pointing the platform at manufacturing, energy, utilities, transportation, logistics and critical infrastructure — sectors where continuous monitoring of large sites can catch equipment failures before they cascade. The companies said they will scale the deployment across Asia-Pacific first, with global availability to follow.
Reporting based on coverage from The Robot Report, Avnet, Weston Robot, ANI News and PR Newswire.