Boston Dynamics' Stretch attacks warehousing's worst job: unloading trucks by hand

Truck unloading was the warehouse's last unautomated job: inside a steel trailer with no power and no infrastructure, in summer heat or winter cold, workers lift boxes all day. It is physically brutal and chronically short-staffed. Boston Dynamics — famous for Atlas and Spot — saw that the industry had automated storage and transport but not the trailer, and in March 2021 unveiled Stretch: a square wheeled base carrying a perception mast and a seven-degree-of-freedom arm with a suction-pad array that lifts boxes up to 23 kg (50 lbs).

Stretch was derived by subtraction. Its ancestor Handle, a two-wheeled research robot spun off from the Atlas team, failed the economics of truck unloading in trials: it maneuvered its whole body around each box, taking about 25 seconds per case against a human's roughly five. "We knew we needed to hit that kind of number for a return on investment," recalls warehouse-robotics head Kevin Blankespoor. Stretch's answer was to use the arm through its full reach and move the base only when needed — combining mobility and manipulation, which the market sold separately but not together.

Customers bought the bet. In January 2022 DHL committed $15 million for Stretch; in August 2022 supply-chain firm NFI added a $10 million order for its US warehouses, with the first unit due in Savannah, Georgia in 2023. Clothing retailers Gap and H&M also signed on. Opening general sales in March 2022, CEO Robert Playter framed Stretch as taking over "one of the most physically demanding jobs in the warehouse" amid labor shortages and supply-chain snags.

Derive a task-specific robot; move the arm, not the base

The trailer was the one place in logistics still worked entirely by human muscle — automation there had no incumbent to displace.

Handle's failure set the key constraint: cycle time, not dexterity. Stretch reaches ROI by working the arm, not driving the base.

Mobility means zero infrastructure: the robot drives into any trailer or container, unlike fixed unloading gantries.

Deriving from Atlas and Handle research let Boston Dynamics ship a product without inventing new core technology.

Research platforms become products by subtraction: strip a general-purpose robot to one punishing task, then fix the single metric — cycle time — that decides whether automation pays.

Stretch became the vanguard of Boston Dynamics' push into mobile manipulation: the company plans for one robot to unload inbound trailers in the morning, build pallets in the afternoon and load outbound trucks in the evening. A fixed sibling, Pick, followed for stationary box handling. As warehouse labor shortages persisted, trailer unloading turned from the industry's least attractive job into its hottest automation frontier, and Boston Dynamics — under Hyundai — positioned Stretch as its first large-scale commercial bet beyond Spot.