GXO, Apptronik plan up to 1,000 Apollo humanoid robots in warehouses

GXO Logistics, the contract logistics company, and the US robot maker Apptronik announced on Thursday, 12 November 2026 that they have signed a letter of intent to expand their work with Apptronik's Apollo humanoid robots. Under the plan, the two companies will run wider trials of Apollo in live warehouse operations from 2027 and build a commercial model in which GXO rents the robots as a service rather than buying them. If the robots meet agreed performance, safety and operational targets, the companies intend to move towards a global roll-out of up to 1,000 Apollo robots over the next four years, starting in North America in late 2027.

Warehouses have used robots for years, but mostly machines built for a single job: shelves on wheels that move to a picker, robotic arms fixed at a packing station, or conveyor-based sorters. Humanoid robots are different. They are built in roughly human shape so that, in principle, they can work in spaces designed for people, take on several different tasks and move between them without the building being redesigned. That is why logistics firms are interested, and also why the claims need testing: a general-purpose robot is far harder to make reliable than a single-purpose one.

The structure of this deal reflects that caution. It is a letter of intent, the roll-out is conditional on the robots meeting agreed performance, safety and operational targets, and the first larger deployment is not due until late 2027. The headline figure of 1,000 robots is an upper limit, not a guaranteed order. GXO's chief executive, Patrick Kelleher, described humanoid robots as "the latest evolution of intelligent automation, and their increasing dexterity will make them one of the most transformative applications of physical AI in logistics." The term "physical AI" refers to artificial intelligence that controls machines acting in the real world, as opposed to software that only produces text or images.

The Robotics-as-a-Service model is also significant for the wider industry. Renting robots by use spreads the cost over time and shifts some of the risk to the robot maker, which must keep the machines working to get paid. If that model proves workable for humanoids, it could make the technology accessible to warehouse operators that could not afford to buy expensive robots outright.

For Indian businesses running large warehouses for e-commerce, retail and manufacturing, the trials will be an early test of whether humanoid robots can do routine warehouse work at commercial scale, and on what terms.

Peepals calculation: A roll-out of up to 1,000 robots over four years works out to an average of up to 250 robots a year (1,000 ÷ 4 = 250). Because the first deployment under the plan is set for late 2027, most of that number would have to come in the later years if the full figure is reached. Our calculation from the reported figures.

Short-term: The two companies will prepare for expanded trials in 2027 across picking, case and tote movement, packing and sortation, and work out the commercial terms of the rental model.

Long-term: If Apollo meets its targets, GXO could become one of the first large-scale users of humanoid robots in logistics, which would encourage other warehouse operators and robot makers. If it does not, the conditional structure lets GXO limit its commitment.

Who is affected: GXO's warehouse operations and staff, its retail and industrial customers, Apptronik and rival humanoid robot makers, and logistics companies weighing automation.

What to watch next: Expanded trials starting in 2027 and the first deployment under the plan in North America in late 2027.

Questions and Answers