Otto Group puts AI robots on its returns line: 25 Sereact systems to grasp, grade and sort what shoppers send back

Otto Group and robotics software company Sereact announced on October 28, 2026 that they will put AI-controlled robots to work on one of retail's most stubborn manual tasks: inspecting and sorting returned customer orders. The plan, in two phases, starts with ten AI-controlled dual-arm systems at Otto Group's logistics center in Poland, with 15 more to follow based on initial results — scaling the program to 25 robots in total.

Returns resist automation for a simple reason: nothing arrives in a predictable state. Items come back folded, crumpled and repackaged, and a multi-category retailer's assortment never stops changing. Traditional warehouse automation depends on known products and predictable presentation, and returns offer neither. Sereact's answer is Cortex, its Physical AI model, which the company says generalizes to objects it has never encountered — no product-specific teaching, no fixed item catalog — and decides in real time how to grasp each item, assess its condition and route it onward.

“Our customers expect returns to be simple and fast – behind that lies a process that is time-critical and laborious, both for customers and for our business. Returns are notoriously hard to automate, and Physical AI has the potential to change that. This project lets us find out, at real scale, how far Physical AI can take us – an important step towards a more flexible and resilient supply chain,” said Martin Umland, Group Vice President Supply Chain Management at Otto Group.

The NVIDIA stack behind the hands

Cortex is built on NVIDIA's physical AI stack. Sereact uses NVIDIA Isaac Sim and Isaac Lab to turn real picks from the returns line into thousands of simulated variants, and NVIDIA Cosmos world foundation models to re-render them across garment states, packaging, materials and lighting. New policies are post-trained on NVIDIA Blackwell GPUs and tested against held-out cases before they reach the robots — so when a robot struggles with an item, the fix can come back as a tested update within days instead of weeks, according to the companies.

The deployment runs as Robotics-as-a-Service: Sereact operates and maintains the systems. People stay in the process, but their work changes — away from repetitive grasping and lifting, toward supervision, exception handling and final quality decisions, the tasks that depend on experience and judgment.

A test at real scale, not a rollout — yet

The wording matters here. This is a joint development project with staged introductions, each generating operational feedback, and Otto Group will only evaluate a broader rollout across its logistics network if the second phase succeeds. The announcement gives no date for either phase and no throughput figures, so the honest read is that the technology has to clear reliability, flexibility and quality bars it has not yet publicly cleared at this site.

The context cuts in the project's favor. Sereact's Cortex is already sorting returns in live operations: a week earlier, on October 21, Sereact, NVIDIA, Nebius, CEVA Logistics and Zalando announced a closed training loop for the same model on Zalando returns handling at CEVA fulfillment centers, with the same days-not-weeks update cycle. Sereact, founded in Stuttgart by Dr. Ralf Gulde and Marc Tuscher, raised a $116 million Series B in 2026 and counts Zalando, IKEA, DHL, BMW and Mercedes-Benz among its customers. Otto Group, meanwhile, frames the project inside a medium-term commitment to invest €350 million in technology and AI.

“A robot that only knows yesterday's catalog is of little use in a returns center. Cortex has to deal with whatever comes out of the next parcel. Every return the systems handle makes the model better at the next one,” said Dr. Ralf Gulde, CEO and Co-Founder of Sereact.

What to watch

Three things will decide whether this becomes more than a pilot: what phase one actually delivers on the returns line, what triggers the additional 15 systems, and whether the robots hold up as the assortment churns through seasons. If they do, Otto Group has a template for automating the messiest corner of e-commerce logistics. If they don't, the project still answers a useful question — how far today's physical AI gets when the catalog is whatever shoppers just sent back.

Sources: Otto Group and Sereact joint announcement via GlobeNewswire (October 28, 2026); Sereact, NVIDIA, Nebius, CEVA Logistics and Zalando announcement via GlobeNewswire (October 21, 2026). The editorial illustrations are AI-generated and do not depict the actual deployment.