Boston Dynamics Rebuilds Stretch Robot With 81 Percent Less Cabling After 100,000 Field Hours

The trailer unloading dock is one of the least glamorous bottlenecks in logistics. A trailer arrives packed floor to ceiling with cases, and every one of them has to come out before the freight can move. For years that work has been manual, repetitive and physically punishing, and the operators of parcel hubs, retail distribution centers and third party logistics networks have spent heavily trying to automate it. Boston Dynamics unveiled Stretch Gen 2 on November 3, 2026, a rebuilt version of its autonomous mobile robot for trailer and container unloading, with the emphasis on keeping the machine working rather than on adding raw speed.

Stretch is not a new product line. The first generation went on sale in March 2022, with DHL Supply Chain as its first commercial customer. AUTONOMOUS reported on November 4, 2026 that Boston Dynamics says Stretch now has more deployments than any competing product in logistics, and that DHL Group signed a memorandum of understanding in May 2025 covering more than 1,000 additional units worldwide. DHL reported unloading rates of up to 700 cases an hour and said it had invested more than 1 billion euros in automation in its contract logistics division over the previous three years. Stretch is no longer a demonstration, it is a fleet.

The redesign began with operating data. Boston Dynamics reported on November 3, 2026 that its team analyzed more than 100,000 operating hours and tens of millions of boxes that first generation machines had already unloaded, then worked backward to find the most common causes of unplanned downtime. The answer was rarely the robot's ability to see a box or plan a pick. It was wiring, connectors, joints striking mechanical stops, and the small physical failures that accumulate when a machine works a dock shift after shift. Gen 2 is the company's attempt to engineer those failures out.

Boston Dynamics is owned by Hyundai Motor Group, and the robot sits in a crowded automation market that includes purpose built unloading systems, mobile manipulators and a wave of humanoid programs. What separates Stretch from most of that field is that it is already deployed, already measured in hours and boxes, and already sold through a conventional commercial channel. The Gen 2 launch is therefore less a technology reveal than a maintenance overhaul informed by a real installed base.

Key Facts

The most consequential change is the wiring. Boston Dynamics cut 81 percent of the robot's internal cabling, roughly half a mile of wire, and described that cabling as the biggest source of intermittent electrical faults and wiring wear. The company replaced dynamic cables and complex wiring harnesses with a simpler architecture, and moved to heavy duty automotive grade connectors rated for the vibration produced by rapid high torque box handling and uneven trailer floors.

The mechanical package received similar treatment. The gripper gained new valve systems and more precise sensing. The joints now rotate continuously beyond 360 degrees, so they no longer strike hard stops during a pick cycle. Standardized high strength fasteners reduce the number of tools a technician needs, and onboard diagnostics let customer maintenance teams handle more routine service. A full swerve drive swap now takes about one hour instead of a full day, the clearest illustration of what Boston Dynamics means by uptime.

Safety changes target interruptions that never show up in a throughput spec sheet. A falling box used to be able to break a panel or trigger a safety stoppage. The emergency stop now sits lower and recessed, and new exterior panel materials make a falling box far less likely to take the robot out of service.

On performance, Boston Dynamics rates Stretch Gen 2 at up to 1,000 cases per hour, a figure that depends heavily on case weight and on how neatly or irregularly a trailer is stacked. A multipick feature lets the robot handle multiple boxes in a single picking cycle for peak rates. The product page lists cases weighing up to 50 pounds. The battery is designed to last two shifts, and a fast charger reaches 90 percent in about 75 minutes. Built in lighting allows work inside dark trailers, and a pre trained vision system detects most packages from day one without programming.

Interesting Engineering reported on November 3, 2026 that the changes target mechanical wear and electrical faults, and quoted Brian Donnelly, Vice President and General Manager of Logistics Robotics at Boston Dynamics, saying that customers count on Stretch to keep freight moving. RoboHub reported on November 4, 2026 that the redesign is aimed at parcel operations, retailers and third party logistics providers, and noted that the performance figures are manufacturer claims rather than independent test results.

Analysis

The bigger picture here is that Stretch Gen 2 is a maintenance product, not a speed product. Boston Dynamics did not lead with a dramatic throughput jump. It led with a half mile of wire removed, a repair that fell from a day to an hour, and a connector designed for vibration. That is an unusual way to launch a robot, and it is the right one. The customers who buy trailer unloaders are not short of demonstrations. They are short of machines that stay running through a peak season.

The economics of dock automation are dominated by availability. A robot that unloads 1,000 cases an hour is useful only while it is actually unloading. Donnelly's framing captures the operational reality: every hour a robot sits idle is an hour of work piling up on the dock. When a swerve drive failure meant a full day of lost work, the cost was not the part, it was the backlog, the overtime and the missed trailer schedule. Compressing that to about an hour changes the math on spare parts, on service contracts and on how many units a site can justify.

What this really means is that the field data loop has become the product. Boston Dynamics studied 100,000 hours and tens of millions of boxes, found that wiring and hard stops caused a disproportionate share of stoppages, and rebuilt the machine around those findings. This is the same logic that turned industrial trucks and aircraft into mature platforms: iterate on failure modes, not on brochures. The redesigned onboard computing platform, which receives regular software releases improving case detection, motion planning, pick strategies and world modeling across a fleet without hardware changes, extends that loop after deployment.

One caveat: the figures are the manufacturer's. RoboHub flagged that the 1,000 cases per hour number is a conditional maximum that depends on case mix and trailer conditions, and that the operating hours behind the redesign come from first generation deployments rather than from an independent study of Gen 2. Buyers should treat the claims as a starting point for a site assessment, not as a guarantee.

Why It Matters

Trailer unloading is one of the last manual strongholds in a warehouse that is otherwise increasingly automated. Palletizing, sortation and storage have been mechanized for years, but a trailer full of mixed cases still depends on people. That makes dock labor expensive, hard to recruit and vulnerable to turnover, and it makes any credible automation attractive to parcel carriers, retailers and logistics providers facing volume peaks they cannot staff.

Serviceability may be the deciding factor in adoption. Early automation often failed not because the robot could not do the task but because every fault required a specialist visit. Onboard diagnostics, standardized fasteners, modular plug and play parts and a one hour drive swap shift routine maintenance to the customer's own team. That lowers the skill barrier, shortens downtime and makes the total cost of ownership easier to model, which is exactly what procurement teams need before they commit to a fleet.

Competitively, the launch also strengthens the company's position in a market where many rivals are still piloting. More than 1,000 additional units covered by the DHL memorandum, and DHL's stated investment of more than 1 billion euros in contract logistics automation, show the scale of demand. A robot backed by a large installed base and a steady software pipeline is a safer bet for a multi site rollout than a newer machine with better headline specifications.

Next Up

Boston Dynamics has not disclosed pricing, a shipping date or which customers will receive the first Gen 2 units, and those details will determine how quickly the redesign reaches the dock. The company says the new computing platform will keep improving case detection, motion planning and world modeling through software releases across the fleet, so existing owners should expect capability gains without hardware changes.

The next real test is field performance. The claims to watch are not the 1,000 cases per hour peak but the mean time between stoppages, the share of repairs handled by customer teams, and whether the one hour drive swap holds up in real service bays. If those numbers hold, Gen 2 will be judged as the moment Stretch became a maintenance mature platform rather than an impressive prototype.