UPS Reportedly Discusses Deploying Humanoid Robots with Figure AI

United Parcel Service (UPS) is reportedly in talks with the California-based robotics startup Figure AI to deploy humanoid robots across parts of its logistics network, according to a report by Bloomberg. 

The talks began last year and have continued into recent months, with the scope of work for the robots yet to be finalised. UPS and Figure have not publicly confirmed specific roles. In February, Figure posted a video showing its robot picking and sorting parcels beside a conveyor belt, suggesting potential logistics tasks.

Our first customer use case took 12 months, our second customer use case took just 30 days

Helix learned high-rate logistics with a single neural network

On Sunday, we successfully validated this on-site at the customer pic.twitter.com/ev5OeSEhly

— Figure (@Figure_robot) February 26, 2025

The move comes as UPS seeks new ways to expand its automation efforts, while Figure AI aims to position itself among the leading developers of humanoid robotics. Startups and major technology firms are increasingly advancing the development of humanoid robots through breakthroughs in AI. 

Figure AI has drawn significant attention in this area and released Helix, a Vision-Language-Action (VLA) model that allows humanoid robots to perform complex tasks using natural language.

A UPS spokesperson told Bloomberg that the company would not comment on “specific or potential vendor partners,” but added, “We regularly explore and deploy a wide range of technologies, including robotics.” 

UPS has already integrated automation technologies into its operations, including robotic arms and AI-powered sorting software at its Velocity facilities. It has also previously partnered with Dexterity, a company producing robots with fine motor skills for warehouse tasks. Additionally, it has explored robotic unloading and autonomous guided vehicles (AGVs) through partnerships with Pickle Robot, Dane Technologies, and other companies.

Meanwhile, Figure recently introduced learned natural walking for its Figure 02 humanoid robot, making it capable of human-like movements. 

This development aimed to enhance the robot’s adaptability for both industrial and domestic applications by mimicking natural human locomotion, thereby compressing years of simulated training into hours.