Amazon unveiled a new version of its autonomous warehouse robot Proteus that workers can direct with spoken language instead of code, the company announced on June 4, 2026. The upgrade strips out the specialized software interface that has fronted Proteus since the robot's 2022 debut and replaces it with a natural-language layer. Amazon plans to deploy the system across Europe in the first half of 2027.
Proteus is the floor-level, tortoise-shaped robot Amazon Robotics uses for heavy lifting and moving large carts through fulfillment centers. Until now, assigning it work meant going through purpose-built software. The new model lets warehouse staff hand off tasks the way they would to a human colleague, with the robot resolving priority, routing, and timing on its own.
Scott Dresser, vice president of Amazon Robotics, framed the shift as moving the cognitive load from operator to machine.
“You tell it what needs to be done. It figures out the priority, the route, the timing.”— Scott Dresser, Vice President of Amazon Robotics
Key facts
- 01Amazon's next-generation Proteus accepts natural-language commands, replacing the specialized software interface used since the robot debuted in 2022.
- 02The upgraded system is being piloted in Amazon Robotics labs with a European deployment scheduled for the first half of 2027.
- 03Earlier Proteus units operated only in dock areas; the new version is designed to move containers anywhere across fulfillment and delivery sites.
- 04Amazon is also expanding its Vulcan touch-sensitive robot and a Barcelona-piloted tote-handling system to more European sites in the coming year.
- 05Amazon says it has hired hundreds of thousands of employees globally since introducing robotics into its operations.
The next-generation Proteus also expands where the robot can work. Earlier units were limited to dock areas, handling containers as they arrived on site. Amazon says the new system can operate anywhere items need to be moved, including transfers between workstations and assistance across fulfillment centers and delivery sites. The hardware design itself has not changed substantially from the 2022 version.
The robot is still in pilot inside Amazon's labs. The Europe-first deployment plan, set for the first half of 2027, sits alongside two other robotics expansions Amazon has on deck for the region: its touch-sensitive robot Vulcan, and a collaborative tote-handling system first piloted in Barcelona. Both are scheduled to reach more European sites in the coming year.
The language interface is the substantive change. Voice and natural-language control of industrial robots has historically tripped over noisy warehouse acoustics, ambiguous instructions, and the cost of error when a 1,000-pound cart goes the wrong way. Amazon's pitch is that a model handling priority and routing inside the robot's own software stack removes the brittle middle layer where human-written task scripts used to live.
The labor framing is the part Amazon is most careful about. The company says it is creating new jobs alongside these technologies and claims to have hired hundreds of thousands of employees globally since introducing robotics into its operations. It insists the robots are designed to support workers rather than replace them, a line it has repeated through every wave of warehouse automation announcements.
That claim runs into a structural question Amazon has not directly answered: as natural-language tasking lowers the skill floor for operating Proteus, and as the robot's working envelope expands from dock zones to entire fulfillment buildings, the marginal human role shifts from supervisor of a specialized system to handler of edge cases. Headcount can grow even as the per-package labor input falls. Amazon has not published a ratio of robots to workers across its fulfillment network, nor a forward projection for either figure.
The Proteus update fits a pattern across Amazon's robotics roadmap of pairing existing hardware with newer model-based control software. Vulcan handles the dexterity problem, the Barcelona tote system handles human-robot handoff, and Proteus now handles the instruction-parsing problem. Each piece is incremental; together they describe a fulfillment center where most physical movement is robot-mediated and most human instruction is voice.
For the broader AI market, the Proteus voice interface is a useful data point on where language models are landing in industrial settings. The deployment is small, the timeline is more than a year out, and Amazon has not named the model stack underneath. But the direction is clear: the company with the largest fulfillment footprint in the world is treating natural-language tasking as the default operator interface for new robotics hardware, not an experimental add-on. Whoever supplies the models that run inside that interface — whether Amazon's own or a partner's — gets one of the densest real-world deployment surfaces in logistics.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
