Amazon in licensing deal with AI robotics start-up Covariant, hires co-founders
Amazon is continuing to grow its AI and robotics team in the Bay Area.
The online giant is hiring three co-founders of Covariant, a start-up that builds AI for advanced warehouse robotics systems, and about one-quarter of its current employees. In addition, Amazon has signed a non-exclusive license to utilize robotic foundation models from Covariant that enable robots to see, reason and act on events happening around them.
According to Amazon, Covariant’s models will help generalize how its robotic systems learn and enable it to use automation to make operations safer and customer delivery faster and more accurate.
Covariant co-founders Pieter Abbeel, Peter Chen, Rocky Duan and a group of the company's research scientists and engineers will join Amazon’s fulfillment technologies and robotics team to help drive the development and implementation of the company’s technology within Amazon. Covariant will continue operating as normal and developing technology that supports fulfillment and distribution center automation.
"We’re excited to welcome some of Covariant’s talented team to Amazon and look forward to building on this technology with them to deliver the next generation of robotics that will drive forward the state-of-the-art in automation," said Joseph Quinlivan, VP, Amazon fulfillment technologies & robotics., in a corporate blog post.
Amazon focuses on next-gen fulfillment center robotics
Amazon has been actively developing new, leading-edge robotic solutions for use in its fulfillment centers. The e-tailer previously said it has over 750,000 robots working collaboratively with its employees, taking on highly repetitive tasks.
[READ MORE: First Look at Amazon's new robotic fulfillment hub in Mass.]
Examples of fulfillment center robot models include the following:
Titan
Amazon leverages a mobile robot called Titan at a fulfillment center in San Antonio. Titan helps carry products across the center including larger, bulkier items like small household appliances or pallets of pet food and gardening equipment.
Digit
In September 2022, Amazon began testing Agility Robotics' bipendal robot, Digit, at its robotics research and development site south of Seattle. These are mobile robots which are built in a human-like shape and can move like a person while also grasping and handling items with robotic "arms" resembling a human.
Sequoia
Building off a series of research and development efforts, Sequoia integrates multiple robot systems to containerize Amazon’s inventory into totes, bringing together mobile robots, gantry systems, robotic arms and a new ergonomic employee workstation.
These totes come to employees at an ergonomic workstation that allows them to do all their work between mid-thigh and mid-chest height. With this system, employees will no longer have to regularly reach above their heads or squat down to pick customer orders, supporting Amazon efforts to reduce the risk of injuries.
Hercules
The proprietary Amazon robot model called Hercules helps warehouse employees by traveling around facilities to retrieve shelves of products and delivering them to employees, who then pick the items customers ordered for shipping. Hercules can lift up to 1,250 pounds and travel across the 1 million sq. ft. of Amazon’s largest fulfillment centers.
Sparrow
In November 2022, Amzazon rolled out Sparrow, an intelligent robotic system that can detect, select, and handle individual products in its inventory. Leveraging computer vision and artificial intelligence, Sparrow can recognize and handle millions of unique items for picking, stowing and packing before they are shipped – at scale.
Proteus
In June 2022, the company announced Proteus, its first fully autonomous mobile robot. Proteus moves autonomously through Amazon’s fulfillment and sort facilities using advanced safety, perception, and navigation technology developed by Amazon.
Cardinal
In June 2022, Amazon also introduced Cardinal, a robotic lifting arm that uses AI and computer vision to efficiently select one package out of a pile of packages, lift it, read the label and precisely place it in a cart.
"With some of the smartest minds, we will advance fundamental research, marrying our rich expertise to unlock new ways for AI and robots to assist our operations employees," said Quinlivan in the blog post. "Embedding Covariant’s AI technology into our existing robot fleet will make them more performant and create real world value for our customers."
