Amazon has quietly shelved one of its most ambitious warehouse robotics projects, discontinuing the humanoid-style robot known internally as “Digit” after determining that the machine simply wasn’t up to the demanding pace of its fulfillment centers. Yet even as the company walks away from one high-profile bet, it is doubling down on automation with a new generation of robots that could reshape how millions of packages move from shelf to shipping dock.
The decision to cancel the Digit program—a collaboration with Agility Robotics, a startup in which Amazon had invested—marks a notable inflection point for the retail and logistics giant. It underscores both the extraordinary difficulty of deploying humanoid robots in real-world industrial settings and Amazon’s willingness to cut losses quickly when a technology fails to meet its exacting operational standards.
What Went Wrong With Digit
Digit, a bipedal robot developed by Agility Robotics, was designed to handle tote-moving tasks inside Amazon’s sprawling warehouse network. The robot could walk on two legs, grasp containers, and theoretically operate in spaces originally built for human workers. Amazon began testing Digit at a facility in the Houston area, and the early vision was tantalizing: a fleet of humanoid machines working alongside people, picking up the most repetitive and physically taxing labor.
But according to reporting by TechRadar, the robot fell short on a metric that matters more than almost anything else in Amazon’s logistics operation—speed. Digit reportedly could not keep up with the throughput demands of a modern fulfillment center, where every second of delay compounds across millions of daily shipments. The robot’s bipedal locomotion, while impressive from an engineering standpoint, proved less efficient than simpler wheeled or tracked alternatives for the specific tasks Amazon needed performed. The company determined that the technology, at least in its current form, was not a viable path to the productivity gains it required.
A Broader Pattern in Humanoid Robotics
Amazon’s retreat from Digit is not an isolated event. It fits into a broader industry reckoning over the practical limits of humanoid robots in commercial and industrial applications. Companies across the technology sector have poured billions of dollars into bipedal and human-shaped machines, driven by the appealing notion that robots built like people could slot into environments designed for people. Tesla’s Optimus project, Figure AI’s partnerships with BMW, and several Chinese robotics firms have all pursued similar visions.
Yet the gap between demonstration-ready prototypes and production-ready industrial tools remains vast. Humanoid robots face fundamental challenges with battery life, reliability, manipulation dexterity, and the sheer computational overhead required to balance and walk while performing useful work. Amazon’s experience with Digit suggests that even with significant investment and a willing deployment partner, these machines are not yet ready for the punishing tempo of high-volume logistics. The lesson for the industry is stark: looking human is not the same as being useful.
Blue Jay Lives On—And New Robots Are Coming
While Digit has been retired, Amazon’s broader robotics ambitions remain very much alive. The company’s internal robotics division continues to develop and deploy a range of automated systems under the umbrella program known as “Blue Jay.” As TechRadar reported, new robots are expected to emerge from the Blue Jay initiative in the near term, with Amazon signaling that it has additional machines in the pipeline that are closer to meeting its performance benchmarks.
Amazon has already deployed tens of thousands of robots across its fulfillment network. The company’s acquisition of Kiva Systems in 2012 for $775 million was an early and defining move, giving Amazon the orange, shelf-carrying robots that became a fixture of its warehouse floors. Since then, the company has introduced robotic arms for sorting, autonomous mobile robots for transporting goods, and sophisticated computer vision systems for inventory management. The cancellation of one project, however prominent, does not change the fundamental trajectory: Amazon is automating its warehouses at an accelerating rate.
The Economics Driving Amazon’s Automation Strategy
The financial logic behind Amazon’s robotics push is straightforward and powerful. The company operates more than 1,000 fulfillment and logistics facilities worldwide and employs roughly 1.5 million people, making it one of the largest private employers on the planet. Labor represents a massive and growing cost, particularly as warehouse wages have risen in recent years amid tight labor markets and increased scrutiny of working conditions. Every robot that can reliably perform a task previously done by a human worker represents a potential long-term cost saving—provided the robot can match or exceed human speed and accuracy.
Amazon’s leadership has been candid about this calculus. In recent earnings calls and public statements, executives have pointed to automation as a key lever for improving margins in the company’s retail and logistics operations. The company spent an estimated $1 billion on robotics research and deployment in 2024 alone, according to industry analysts. That investment is expected to grow as new systems prove their value and as the technology matures. The Digit setback is a speed bump, not a roadblock.
What This Means for Amazon’s Workforce
The acceleration of warehouse automation raises pointed questions about the future of Amazon’s massive hourly workforce. The company has consistently maintained that robots are meant to complement human workers, not replace them—a claim that labor advocates and economists view with varying degrees of skepticism. Amazon has pointed to the fact that its total headcount has grown even as it has deployed more robots, arguing that automation creates new categories of jobs, such as robot technicians and systems operators.
But the long-term trend is harder to dismiss. As robots become more capable and less expensive, the economic incentive to reduce reliance on human labor intensifies. Warehouse workers performing repetitive tasks like moving totes, stacking boxes, and sorting packages are the most immediately vulnerable. Unions and labor organizations have raised alarms about the pace of automation, arguing that workers deserve a voice in how these transitions are managed and that the gains from automation should be shared more broadly.
The Competitive Pressure From Rivals
Amazon is not the only major logistics player investing heavily in warehouse automation. Walmart, its chief retail rival, has partnered with Symbotic to deploy automated systems across its distribution centers. FedEx and UPS have both increased their robotics spending. In China, companies like JD.com and Cainiao (Alibaba’s logistics arm) have built highly automated facilities that in some cases operate with minimal human intervention. The competitive pressure to automate is global and intensifying.
This context helps explain why Amazon is willing to absorb the cost of a failed project like Digit and move on quickly. The company cannot afford to fall behind in automation, because doing so would erode the speed and cost advantages that underpin its dominance in e-commerce fulfillment. Every competitor that deploys a faster, cheaper, or more reliable robotic system puts pressure on Amazon to match or exceed that capability.
The Road Ahead for Amazon Robotics
Looking forward, Amazon’s robotics strategy appears to be shifting toward purpose-built machines rather than general-purpose humanoid platforms. The company’s most successful deployments have involved robots designed for specific, well-defined tasks: moving shelves, sorting packages, palletizing goods. These machines may lack the visual drama of a walking humanoid, but they deliver measurable productivity improvements in controlled environments.
The Blue Jay program is expected to yield new systems that follow this philosophy—machines optimized for particular warehouse functions rather than attempting to replicate the full range of human physical capability. Amazon’s internal robotics team, which includes engineers from the Kiva acquisition and subsequent hires from across the robotics industry, has the resources and institutional knowledge to iterate rapidly. The company has also signaled interest in AI-driven improvements to its existing robot fleet, using machine learning to optimize routing, picking, and placement decisions in real time.
Amazon’s decision to cancel the Digit project is a reminder that even the world’s most powerful technology companies face real limits when pushing the boundaries of automation. But it is also a signal that Amazon’s commitment to replacing manual warehouse labor with machines is unwavering. The robots may not walk on two legs, but they are coming—faster and in greater numbers than ever before.