In 2012, Amazon quietly acquired a robotics startup called Kiva Systems.a significant improvement the effectiveness of its e-commerce operations and kickstarted a broader revolution in warehouse automation.
Last week, the e-commerce giant announced another deal that could prove similarly profound, agreeing Covariant Founders HireA startup has been experimenting with ways to use AI to automate the picking and handling of a variety of physical objects.
Covariant may find commercializing AI-integrated industrial robots a challenge due to high costs and fierce competition; this deal, in which Amazon also licenses Covariant’s models and data, could usher in another revolution in e-commerce — one that may be difficult for any competitor to match given Amazon’s scale of operations and massive data stores.
The deal is also an example of a major tech company acquiring core talent and expertise from an AI startup without actually buying the company outright. Amazon has come a similar agreement with startup Adept in June. In March, Microsoft acquired Dealing with Inflectionand in August, Google hired the founders of Character AI.
Back in the 2000s, Kiva developed a way to move products through warehouses by having squatting robots lift and carry filled shelves to pickers—a trick that saved workers from having to walk miles each day to find different items. Kiva’s mobile robots are similar to those used in manufacturing, and the company used smart algorithms to coordinate the movements of thousands of robots in the same physical space.
Amazon’s mobile robot army has grown from around 10,000 in 2013 to 750,000 in 2023, and the company’s massive scale of operations means it can deliver millions of items faster and cheaper than anyone else.
EQUAL WIRED revealed last yearAmazon has in recent years developed new robotic systems that rely on machine learning to perform tasks like recognizing, picking, and sorting packaged boxes. Again, Amazon is using scale to its advantage, with training data collected as items flow through its facilities helping to improve the performance of various algorithms. This effort has led to further automation of work previously done by workers at some fulfillment centers.
One task that remains difficult to mechanize, however, is the manual handling of products. It requires adaptability to account for things like friction and slippage, and robots will inevitably encounter unfamiliar and awkward items in Amazon’s massive warehouses.
Covariant has spent the past few years developing more generalizable AI algorithms that can handle multiple items more reliably. The company was founded in 2020 by Pieter Abbeela professor at UC Berkeley did Pioneering work in applying machine learning to roboticsalong with several of his protégés, including Peter Chen, who became CEO of Covariant, and Rocky Duan, the company’s CTO. This week’s deal will see all three, along with several research scientists at the startup, join Amazon.
“Covariant’s models will be used to power a number of robotic manipulation systems across our fulfillment network,” Amazon spokesperson Alexandra Miller told WIRED. The tech giant declined to disclose financial details about the deal.
Abbeel was one of OpenAI’s first employees, and his company was inspired by ChatGPT’s success story. In March, Covariant demonstrated a conversational interface for its robot and said it has developed a platform model to be grasped by robots, meaning an algorithm designed to be
