Sereact and NVIDIA Launch Robotic Returns Workflow
A new AI-driven training loop for robotic returns is accelerating policy updates at CEVA Logistics centers.
Updated on Oct. 21, 2026 in Robotics
Live Poll
Do you believe the expanded use of AI-driven robotics in warehouses will benefit shoppers?
Sereact, NVIDIA, Nebius, CEVA Logistics, and Zalando have unveiled a closed training loop designed to automate complex warehouse returns tasks. The system integrates robot vision and simulation to streamline the processing of highly variable items.
Why it matters
Returns processing is traditionally difficult to automate due to the unpredictable nature of item conditions and packaging states. This new workflow reduces the time required for robot policy updates from weeks to days by utilizing advanced simulation and cloud-based AI.
The system utilizes NVIDIA Isaac Sim and Cosmos models to retrain Sereact Cortex robot brains within Nebius data centers using Blackwell GPUs. This pipeline allows for policy iteration at significantly increased speeds.
The players
Sereact
A Stuttgart-based company specializing in AI for robotics, providing the Cortex robot brain and hardware.
NVIDIA
A multinational technology corporation known for its graphics processing units and advanced AI simulation platforms.
CEVA Logistics
A major global logistics and supply chain management company based in Marseille.
Zalando
A Berlin-based European e-commerce giant serving 62 million active customers and thousands of brands.
Nebius
An Amsterdam-based cloud platform provider focused on high-performance AI infrastructure and data storage.
The details
The pipeline captures episodes from robotic stations, which are then processed through simulation to generate variants for testing and retraining. The solution is now available as a reference architecture on the Nebius AI cloud for industrial application.
Timeline
October 21, 2026: Launch of the closed training loop for robotic returns
The Tech Race
This integration follows the growing industry trend of using the NVIDIA Isaac Sim simulation environment to bridge the gap between digital training and physical warehouse deployment. It positions these firms at the forefront of the race to replace legacy, manual-heavy warehouse processes with autonomous Physical AI.
While the technology operates in the background, consumers may eventually experience faster processing times for returned goods and more consistent order fulfillment. The shift reflects a broader move toward using advanced automation to manage the high volume of retail returns.
The takeaway
The deployment highlights how collaborative AI ecosystems are successfully tackling the variability of real-world logistics tasks. By shifting from manual oversight to automated training, companies can adapt their robotic workforces to new challenges in near real-time.
Further reading
Learn more about the latest innovations in Robotics.
Live Poll
Do you believe the expanded use of AI-driven robotics in warehouses will benefit shoppers?
