In 2012, Amazon quietly acquired a robotics startup called Kiva Systemsa move that dramatically improved the efficiency of its e-commerce operations and kicked off a broader revolution in warehouse automation.
Last week, the e-commerce giant announced another deal that could prove equally profound, agreeing to Hire the founders of Covarianta startup that has been testing ways for AI to further automate the selection and manipulation of a wide range of physical objects.
Covariant may have struggled to market AI-powered industrial robots given the high costs and stiff competition involved; the deal, which will also see Amazon license Covariant’s models and data, could spark another revolution in e-commerce — one that could prove difficult for any competitor to match given Amazon’s massive operational scale and trove of data.
The deal is also an example of a large tech company acquiring key talent and expertise from an AI startup without buying the company outright. Amazon came to A similar agreement with startup Adept in June. In March, Microsoft reached an agreement dealing with Inflectionand in August, Google Hired the founders of Character AI.
In the 2000s, Kiva developed a way to move products through warehouses using crouching robots that lifted and carried full shelves to human pickers — a trick that meant workers no longer had to walk miles every day to find different items. Kiva’s mobile robots were similar to those used in manufacturing, and the company used smart algorithms to coordinate the movement of thousands of robots in the same physical space.
Amazon’s army of mobile robots grew from around 10,000 in 2013 to 750,000 in 2023, and the sheer scale of the company’s operations meant it could deliver millions of items faster and cheaper than anyone else.
As WIRED revealed last year, Amazon has in recent years developed new robotic systems that rely on machine learning to do things like sense, grasp, and sort packed boxes. Once again, Amazon is leveraging scale to its advantage, as training data is collected as items flow through its facilities, helping to improve the performance of different algorithms. The effort has already led to increased automation of work previously done by human workers at some fulfillment centers.
The one task that remains difficult to mechanize, however, is physically picking up products. It requires adaptability to account for factors like friction and slippage, and robots will inevitably be faced with unfamiliar and unwieldy items among Amazon’s vast inventory.
Covariant has spent the last few years developing AI algorithms with a more general capability to handle a variety of elements more reliably. The company was founded in 2020 by Pieter Abbeela professor at the University of California at Berkeley who has conducted Pioneering work in the application of machine learning to roboticsalong with several of his students, including Peter Chen, who became Covariant’s chief executive, and Rocky Duan, the company’s chief technology officer. This week’s deal will see all three, along with several of the startup’s research scientists, join Amazon.
“Covariant’s models will be used to power some of the robotic handling systems in our logistics network,” Amazon spokesperson Alexandra Miller told WIRED. The tech giant declined to disclose financial details of the deal.
Abbeel was one of OpenAI’s earliest employees, and his company has been inspired by the ChatGPT success story. In March, Covariant demonstrated a chat interface for its robot and said it had developed a base model for robotic grasping – that is, an algorithm designed to become a robot.