Amazon has just completed the acquisition of Rightbot Technologies, a startup that develops truck unloading robots. This operation aims to complete its logistics tooling with a very targeted technological brick, in this case, the unloading of trucks and containers, one of the most complex environments that industrial robotics faces.
The operation, the financial terms of which have not been made public, is a continuation of the $6 million investment led by Amazon via its Industrial Innovation Fund in 2023. Rightbot’s teams are now joining the Robotics Delivery and Packaging Innovation (RDPI) division, responsible for designing and industrializing the robotic systems deployed at the heart of the group’s logistics network.
Unloading trucks, an unsolved technological problem
Unlike automated warehouses, where flows are largely standardized, truck trailers concentrate a significant combination of constraints: heterogeneity of packages, unstable stacking, unpredictable geometries, low visibility and absence of repeatable trajectories.
For each package extraction, the robot cannot follow a predefined plan and must decide, in each cycle, which object to remove, where to apply grip and how to adjust its movement based on the reaction of the environment. This complexity explains why unloading remains, despite a decade of innovations, largely dependent on human labor.
Aspiration as an industrial compromise
The solution developed by Rightbot is based on suction cup gripping, which has now become a de facto standard for this type of use.
Suction offers high tolerance to shape and surface variations, reduces mechanical complexity and allows rapid cycles, while limiting the risk of package damage. This approach is also the one adopted by Boston Dynamics with its Stretch robot, or by Pickle Robotalready deployed on an industrial scale.
The main difficulty lies in the perception and interpretation of unstructured scenes. The robot must identify exploitable surfaces, estimate the stability of a stack and anticipate the mechanical effects of each extraction.
The systems developed by Rightbot combine RGB-D vision, segmentation of partially visible objects and feedback loops integrating pressure, force and movement. Dealing with failure is an integral part of operation, so a failed shot, warped cardboard or slip triggers immediate replanning.
In this context, AI is used as a constrained decision engine, intended to maintain a level of reliability compatible with continuous industrial operations.
Why Amazon is internalizing this technology
For Amazon, unloading trucks is not a marginal case. It constitutes a critical crossing point in the logistics chain, where productivity gains are directly correlated to operator safety and fluidity of flows.
By internalizing this technological brick, the group gives itself the capacity to train and optimize it using volumes of data and real scenarios inaccessible to an independent player. Integration within RDPI also makes it possible to couple these robotic systems with orchestration, planning and supervision tools already deployed across the Amazon network.
One of Amazon’s objectives is to reduce dependence on external suppliers and strengthen the group’s operational advantage.
