Trailer and container unloading robots in the UK: what exists, what the makers publish, and what to test

Yes, for floor-loaded cases, and the makers now name customers doing it. Boston Dynamics offers Stretch for “unloading trailers and containers”, states it handles cases “up to 50 pounds” and can “move hundreds of cases an hour”, and lists DHL, Maersk, Gap Inc., H&M, NFI, Arvato and Otto Group on its product page. Pickle Robot states its system “reliably unloads trailers and containers of non-palletized goods in as little as 90 minutes” and names UPS. Humanoid makers position unloading as a target task but, as read on 22 September 2026, publish no unloading rate for any humanoid. So the honest 2026 picture is: mobile-arm unloaders are in customer operations, humanoids are in pilots, and neither has a published UK reference on its maker’s page.

What none of them do is unload a palletised trailer. Pallets come off with a forklift or a pallet mover, and that is a solved problem. The robots on this page exist because a large share of inbound freight, especially deep-sea containers from Asia, arrives floor-loaded to maximise cube, and every one of those cases is currently lifted by a person.

Because it is the worst job in the building and the hardest to staff. A 40-foot container of floor-loaded cases can hold many hundreds of cartons, handled one at a time, from floor level to above head height, at the far end of a box with no ventilation. In the UK that work sits under the Manual Handling Operations Regulations 1992, which require the employer to avoid hazardous manual handling where reasonably practicable and to reduce the risk where it cannot be avoided. The Health and Safety Executive’s guidance treats repetitive lifting of this kind as a leading cause of musculoskeletal injury. Automating it is therefore a safety case as much as a productivity one, which is unusual for warehouse automation and is why finance and health and safety tend to agree on it.

It is also a bounded task with a clear measure. A trailer is empty or it is not; the cases came off undamaged or they did not; it took a certain number of minutes. Few warehouse tasks are that easy to score, which makes unloading a good first pilot even before the robot is proven.

Mobile-arm unloaders

A robotic arm with a vacuum or gripper end effector on a mobile base, driven into the trailer, picking cases from the wall of freight and placing them onto an extendable conveyor behind it. Boston Dynamics’ Stretch and Pickle Robot’s system are the named examples. This is the form with customers in production and published rates, and the form that changes the dock the most, because it needs a conveyor path from the trailer into the building.

Humanoids

A bipedal robot that walks into the trailer and does what a person does: lifts a case and places it on a conveyor or a pallet. The appeal is that nothing about the dock has to change, and the same robot can do other handling work when no trailer is in. The cost is that no humanoid maker publishes an unloading rate, a sustained case weight or an endurance figure for this task, and the machines in the class are in pilots rather than production for it. Our main humanoid warehouse guide sets out what the makers do publish.

Conveyor-based and fixed unloaders

Telescopic boom conveyors that extend into the trailer so a person only lifts each case a short distance, and fully automated unloaders that combine a boom with a picking head. The telescopic boom is the mature, inexpensive half-step most UK docks already know; it reduces the walk and the carry but not the lift. Fully automated fixed unloaders are a capital project for a dock that runs the same freight profile all day.

Five things, and most of them are about the dock rather than the robot.

If the goal is to stop people lifting cases this year, pilot a mobile-arm unloader, because it is the form with production customers and published rates, and accept that the dock will change to suit it. If the goal is to learn what a general-purpose robot can do across the operation, and unloading is one of several handling tasks on the list, a humanoid pilot is the honest experiment, with the expectation that it will be slower than a person at first and that the maker will not give you a rate up front. The two are not exclusive; they answer different questions, and the humanoid-versus-AMR distinction applies here in the same way.

What matters more than the form is the measurement. Agree before the robot arrives what a trailer of your typical freight takes a two-person crew today, in minutes and in damage, and score the machine against that with the same freight. That is the pilot design we use, on a manufacturer-neutral basis, through a structured pilot programme, and where a dock wants an outcome rather than an asset we scope subscription and hire options alongside purchase.