Inside Amazon’s robot revolution: How Sparrow, Proteus, and Cardinal are reshaping warehouse work

Amazon’s warehouse automation revolution

Amazon’s robot story started back in 2012, when it bought Kiva Systems, a company that built squat orange robots to move goods across warehouses. That was the foundation for what’s now Amazon Robotics, which classifies its warehouse tasks into six areas: movement, manipulation, sorting, storage, identification, and packing. Over the years, Amazon has rolled out multiple robot families, each fine-tuned for a specific role—Hercules for moving heavy carts, Pegasus for sorting orders, Robin and Sparrow for picking and placing items. Together, they’ve helped Amazon reduce manual handling while increasing speed and precision, the New York Times reported.

Meet the robots taking over the warehouse floor

At the company’s most advanced facility in Shreveport, Louisiana, these robots now handle almost every stage of fulfilment. Products arrive in plastic bins that slide in and out of frames in a system known as Sequoia. The Sparrow robotic arm then scans the bins using computer vision and suction cups to pick up items and move them to other containers. Robin takes the packed orders and places them on small shuttle bots like Pegasus, which send parcels down chutes. Then, a tall arm named Cardinal stacks boxes neatly into carts—like a game of high-speed Tetris—before a tortoise-like robot called Proteus glides underneath, carrying the loaded carts to the shipping dock.

Why Amazon is racing to automate

The company says the goal isn’t to eliminate humans but to “make work safer and more efficient.” Yet the underlying reality is clear: automation could help Amazon avoid hiring hundreds of thousands of workers in the coming years. By cutting down manual steps—like lifting, sorting, or packaging—Amazon can process more orders with fewer people. The company’s head of operations, Udit Madan, told The New York Times that redesigning warehouse systems like Sequoia helps “simplify processes and improve safety while boosting output.”

What’s still too hard for robots

Even as automation advances, not everything has been conquered. The process of decanting—cutting open boxes and moving items into storage bins—remains largely manual. Amazon’s prototypes for automated decanting haven’t yet matched the speed or care of human workers. For now, people still oversee that step to make sure products aren’t damaged and that inventory matches shipments. In smaller, same-day delivery centres, Amazon is testing a system called Jupiter to improve automated storage and retrieval, but it’s still years away from “lights-out” warehouses that run almost entirely on machines.

The next generation: robots that think

Amazon recently made a $400 million investment to bring in the founding team of Covariant, an AI start-up that builds systems acting as a robot’s “brain.” This technology helps robots like Sparrow “see” better—identifying different shapes, textures, and ideal ways to grip objects. Amazon is also developing new prototypes named Bluejay and Starling that aim to handle a wider range of items across more types of facilities. Together, these systems signal a major leap toward warehouses that blend robotics with artificial intelligence to manage everything from groceries to electronics.

What this means for the future of work

For Amazon, the promise is speed, efficiency, and precision. For workers, it’s a shift in roles—from physically moving boxes to overseeing, maintaining, and training robots. While some jobs will inevitably fade, others requiring technical skills may grow. But as the company pushes toward “near lights-out” automation, the broader question remains: how far can robots go before they completely redefine what it means to work at Amazon?