GXO and Apptronik expand Apollo warehouse trials, consider deployment of up to 1,000 units

GXO and Apptronik signed an LOI to expand warehouse trials of Apollo, their humanoid robot. Beginning in 2027, the companies will evaluate the robot across multiple warehouse tasks against agreed productivity, safety and operational KPIs. Subject to the trial results and meeting those KPIs, they intend to deploy up to 1,000 units globally over the next four years, with an initial North American deployment planned for the second half of 2027.

GXO and Apptronik outlined a plan to tie deployment of up to 1,000 units to trial results and achievement of KPIs, with evaluations beginning in 2027 across multiple processes, including picking and packing. For warehouse operators, the framework means that deployment decisions will depend not only on demonstrations of individual tasks but also on safety, productivity and operating conditions when robots work alongside existing automation and employees.

The figure of up to 1,000 units is not a confirmed deployment number; it reflects an intention to deploy, contingent on successful trials and achievement of agreed KPIs. The central point of the announcement is therefore not the number of units itself, but a framework for measuring safety, productivity and operating conditions across multiple warehouse processes, then moving forward with applications that meet the criteria. An initial North American deployment is also planned for the second half of 2027, but the timing and scale of a global rollout have not been finalized. The evaluation will cover not only picking but also moving cases and totes, packing and sorting. This suggests that Apollo is being assessed not as a machine for a single task, but on whether it can work alongside existing automation and employees across different processes. The proof of concept launched in 2024, a U.S. pilot focused on autonomous picking of small consumer goods, and evaluations of more than 50 tasks at Robot Park show that the expanded trials build on earlier testing. However, task evaluations and sustained warehouse operations or KPI achievement are not equivalent results. The inclusion of RaaS in the commercial framework indicates an approach that would structure robot deployment as a service, rather than solely as an outright purchase of the hardware. This connects with the service-model issues covered in an article on leaderobot.com about Optimus service models and robot rentals, but that article's publication does not establish the economics of this deal. GXO and Apptronik have not disclosed pricing, maintenance terms, utilization rates or measured results by task. The cost structure of RaaS and how many of the potential 1,000 units would actually be assigned to which processes remain uncertain. Forbes' report on GXO's plan to deploy Apptronik humanoid robots and GXO and Apptronik's expansion of Apollo warehouse trials, with a potential deployment of up to 1,000 units also cover the same announcement, but neither reporting nor an LOI confirms a commercial deployment. The next points to watch are process-specific KPIs and their measured results, the actual number of units deployed, and operating and maintenance terms.

The trials will cover goods-to-person and person-to-goods workflows, moving cases and totes, packing and sorting. The companies will also evaluate how Apollo works alongside existing automation equipment and GXO employees. The commercial framework envisions RaaS (Robotics-as-a-Service), which would allow flexible service delivery. The companies began a proof of concept in warehouses in 2024. They have said that a pilot at a GXO site in the U.S. focused on autonomous picking of small consumer goods, while Apptronik's training facility, Robot Park, collected data and evaluated autonomous operation across more than 50 logistics tasks.