Amazon’s Delivery Robots Are “Jogging” to Doors.

Amazon’s Delivery Robots Are “Jogging” to Doors

1. Origins: Enter Scout

• Launch & Deployment
• Amazon unveiled Scout, a six-wheeled autonomous delivery robot, in January 2019. Designed to travel at a walking pace on sidewalks, its first deployment was in Snohomish County, Washington, followed by trials in Irvine, California; Atlanta, Georgia; and Franklin, Tennessee
• Public Trials & Human Operators
• Scout’s rollout included human “ambassadors” who monitored operations and stepped in when needed, ensuring safety around pedestrians, pets, and obstacles

2. How Scout Works

• Navigation & Perception
• It uses sensor arrays (vision, lidar) combined with machine learning to detect curbs, foliage, people, pets, rain, and debris. When GPS falls short, Scout uses precise sidewalk-level maps to decide, for example, “Can I pass between this trash bin and hedge?”
• City & Sidewalk Challenges
• Unlike structured road environments, sidewalks are unpredictable—cluttered with tools, outdoor furniture, children’s toys. This complexity demands advanced algorithms to dodge, reroute, and engage only when levels of pedestrian activity are lower
• Simulation & Training
• Amazon scientists employ rich 3D environments (“digital sandboxes”) to simulate weather, lighting, airborne particles, and seasonal foliage. Scout is trained virtually, then tested physically to master “micro-environments,” ensuring robustness before real-world deployment

3. Scout’s Progress & Pause

• Gradual Expansion
• The Scout project scaled to four test regions in the U.S. between 2019 and 2020
• In October 2022, Amazon paused Scout's public tests, reassigning many employees, but maintaining core efforts to refine the tech
• End of the Line
• Officially canceled in January 2023, Scout was shelved from field tests, although internal development quietly persisted .

4. Why Scout Didn’t Become the Norm

• High Operational & Infrastructure Costs
• Last-mile delivery with Scout required intensive support—including human monitors, sensor maintenance, precise mapping, and regulatory logistics
• Regulatory & Social Barriers
• Local regulations often prohibited sidewalk robots. Even when allowed, public reception ranged from enthusiastic to cautious, with some novelty confrontations like kids petting the robots or occasional vandalism .
• Technology Limitations
• Scout couldn’t climb steps, open gates, or detect many unexpected obstacles in real-time. Human adaptability in last-mile tasks remained unmatched .

5. Enter the Next Phase: Humanoid Robots “Jogging” to Doors

• Humanoid Delivery Bots in Development
• According to recent reports by The Information, Amazon is working on humanoid robots designed to “spring out” of its Rivian electric vans and autonomously handle door-to-door delivery tasks
• The “Humanoid Park” Testing Ground
• A specialized indoor arena in San Francisco, dubbed “humanoid park,” is being used to train these bipedal bots on obstacle navigation (e.g. door opening, mailbox handling), staging them under realistic use cases (
• Integration with Electric Vans
• These robots are being designed to deploy from Amazon’s Rivian fleet—starting with over 20,000 vans, with plans to scale to 100,000 by end of decade. The robots would emerge, jog to your doorstep, deliver, and re-enter the vans

6. Technological & Operational Challenges

• Complex Physical Interactions
• Tasks like gate-opening, climbing steps, operating doorknobs, and placing packages require dexterous manipulators and dynamic balance control—significantly more complex than wheeled systems .
• Highly Controlled Environments
• Trials are occurring in structured settings first; as experts warn, success hinges on restricting bots to similar conditions: clear driveways, standardized doors, minimal clutter
• Deployment of humanoid robots will amplify regulatory scrutiny. Indoor/outdoor transitions, surveillance concerns, liability issues, privacy, and public acceptance all need addressing.

7. Why Amazon Is Doubling Down

1. Technological Leap
2. Advances in AI, combining bipedal mobility, grasping, and vision systems, make humanoid delivery more viable and capable than sidewalk robots ever were.
3. Operational Flexibility
4. Humanoids can work alongside humans—multi-tasking with drivers in vans leads to faster, more efficient drop-offs.
5. Economic Incentive
6. Automating door-to-door delivery could substantially cut labor and logistical costs at scale, addressing the “last mile” pain point in logistics.
7. Sustainability & Brand Value
8. Coupled with Rivian EVs, humanoid bots align with Amazon’s “net-zero by 2040” sustainability promise and reinforce its tech-forward brand identity.

(YouTube clip highlights reported humanoid testing: robot emerges from van, navigates obstacles)

Amazon, the e-commerce and logistics titan known for its relentless innovation, has been exploring robotic delivery solutions for years, and its latest push into humanoid delivery robots—described by insiders as “jogging to doors”—marks a striking evolution in its last-mile delivery strategy. The journey began with the introduction of Amazon Scout in early 2019, a compact, six-wheeled autonomous delivery robot designed to travel at walking speed along sidewalks and deliver small packages directly to customers' homes. Scout operated using a combination of GPS, advanced sensors, and computer vision, allowing it to avoid obstacles like pedestrians, pets, bicycles, trash bins, and unpredictable weather conditions. Amazon initially deployed Scout in Snohomish County, Washington, before expanding to Irvine, California, and several other U.S. cities. While the device garnered public interest for its adorable design and novelty, it never moved beyond pilot testing due to significant limitations. These included high operating costs, the need for human “ambassadors” to accompany the robot during trials, local regulations restricting sidewalk robotics, and the robot's inability to navigate stairs, gates, and cluttered environments common in suburban and urban areas. Amazon officially paused the Scout program in October 2022, reassigning employees and scaling back the initiative, which was eventually shelved entirely in 2023. Despite this, Amazon didn’t abandon the dream of autonomous delivery; instead, it pivoted toward a more ambitious vision involving humanoid robots capable of navigating more complex environments than Scout ever could. These new robots, currently being tested in a private facility dubbed the "humanoid park" in San Francisco, represent a significant technological leap. Unlike Scout, which was restricted to relatively flat and simple paths, the humanoid bots are being designed to step out of Rivian electric delivery vans, jog to customers’ doors, open gates, climb stairs, place packages at specified drop points, and return—all without human intervention. These robots use advanced AI to process real-time environmental inputs, computer vision for precise object recognition, and mechanical limbs that mimic human locomotion, enabling them to perform complex physical tasks like opening doors and pressing doorbells. Amazon’s integration of these robots with its growing fleet of Rivian electric vans—of which more than 20,000 are already deployed—further demonstrates the company’s commitment to sustainable and technologically sophisticated logistics solutions. The decision to go humanoid is rooted in practical advantages: while wheeled robots like Scout could be obstructed by the simplest terrain variations, humanoid robots, with their bipedal form, can handle many of the physical challenges humans face, making them far more flexible for real-world conditions. Additionally, humanoid robots don’t require the same kind of special sidewalk infrastructure or permissions, potentially allowing smoother regulatory integration in many areas. Nevertheless, this next step in robotic delivery is not without its hurdles. The technology required to make these robots function reliably in varied, uncontrolled outdoor environments is immense; the robots must handle unpredictable events such as a sudden dog barking, children playing, poor weather, or slippery surfaces. Moreover, the robots must respect human privacy, avoid any intrusive surveillance, and be able to perform tasks safely and effectively in highly diverse home layouts. Regulatory bodies may demand extensive testing, safety certifications, and assurance that these autonomous systems won’t malfunction in public or private spaces. Amazon is reportedly training these robots in controlled simulations and real-world facsimiles—recreating everything from doorsteps to fences and walkways—to ensure they are prepared for the nuances of residential environments. But beyond technical and regulatory concerns lies another major factor: public acceptance. Will people be comfortable with a humanoid robot delivering their packages? Will it be seen as helpful or invasive? Amazon seems to believe that the benefits—such as faster delivery, reduced carbon emissions through electric integration, and significant labor cost savings—will eventually outweigh the challenges. The company is positioning itself to lead the race in solving the “last mile” problem, which is the most costly and complex part of delivery logistics. If successful, humanoid delivery bots could eliminate the need for drivers to exit their vehicles, enabling delivery vans to operate more efficiently. Moreover, Amazon is leveraging its extensive logistics network, robotics expertise, and cloud infrastructure to refine and deploy these bots at scale. The move also aligns with the company’s broader sustainability goals, including a commitment to reach net-zero carbon emissions by 2040. Despite the risks, Amazon’s strategy appears well-calculated: instead of investing further in constrained technologies like sidewalk robots that can’t meet real-world demands, it’s focusing on next-gen robotic platforms that could one day be as common as delivery vans themselves. These humanoid bots are being trained to be adaptable, intelligent, and human-aware, capable of interpreting gestures, recognizing faces, and even navigating around curious pets and children. Early prototypes have already demonstrated promising levels of mobility and object manipulation, although mass deployment is still years away. It’s also worth noting that Amazon’s experiments in robotics don’t exist in a vacuum. Competitors like Tesla (with the Optimus robot), Boston Dynamics (with Atlas and Spot), and companies like Agility Robotics and Figure AI are all racing to build commercially viable humanoid robots for industrial and consumer applications. In this race, Amazon’s advantage lies in its colossal delivery volume, deep AI infrastructure (via AWS), and ability to scale hardware manufacturing rapidly through its partners. Should these humanoid delivery bots succeed, they could signal a fundamental shift in how packages are delivered and spark a larger transformation across logistics, retail, and even public infrastructure planning. Over time, neighborhood layouts, drop zones, and household designs might evolve to accommodate robotic delivery norms. As Amazon continues testing and refining these bots, the sight of a jogging robot approaching your front porch may go from science fiction to daily routine. Whether society is ready for this future remains an open question, but one thing is clear: Amazon’s robots are not just walking—they’re jogging confidently toward the next frontier of delivery.

In a bold leap toward redefining last-mile logistics, Amazon has turned the page from its short-lived sidewalk robot, Scout, to a more ambitious and futuristic solution: humanoid delivery robots that are, quite literally, “jogging” to customers’ doors. Originally launched in 2019, Amazon Scout was a small, six-wheeled, cooler-sized autonomous delivery device designed to transport small packages to homes by rolling along sidewalks at walking pace. Using GPS, advanced sensors, machine learning, and real-time obstacle detection, Scout was meant to usher in a new age of autonomous ground delivery, especially in suburban neighborhoods. However, while it showed promise, the limitations quickly became apparent. Scout struggled with stairs, gates, high curbs, cluttered walkways, and the unpredictable nature of real-world sidewalks filled with trash bins, pets, children, uneven pavement, and various unstructured obstacles. It required constant monitoring by human “ambassadors” who accompanied the robot in case of emergency, and public reception was mixed—some found it charming, others saw it as intrusive or gimmicky. Regulatory hurdles also played a significant role; many municipalities lacked clear frameworks for sidewalk robots, or outright banned them. By 2022, despite trials in several cities and thousands of test deliveries, Amazon scaled back the project, ultimately shuttering Scout in early 2023, citing the need to rethink its approach to autonomous delivery technology. But Amazon wasn’t stepping away from automation—it was stepping up. Behind the scenes, the company began investing heavily in a more capable, human-like solution: autonomous humanoid robots that can not only walk and jog, but also climb stairs, open gates, identify safe drop-off points, and interact with human environments with a level of dexterity that Scout could never achieve. These robots are currently being developed and tested in a specialized indoor training area in San Francisco, referred to as “humanoid park,” where Amazon engineers simulate the variety of real-world conditions the bots will eventually face, including doorsteps, porches, delivery boxes, dogs, bikes, and more. Amazon’s vision is to have these humanoid robots deployed from the back of its Rivian electric delivery vans, jog to the customer’s front door with a package in hand, place it safely in the designated area (like a porch or locker), and return to the van—repeating this process with the kind of efficiency, accuracy, and safety that would revolutionize the delivery process. Technologically, this requires an advanced combination of AI, computer vision, robotics, and real-time motion planning. The robots must be able to interpret their surroundings instantly, detect and avoid obstacles, operate in various weather conditions, and perform complex physical actions like walking on gravel, turning doorknobs, pressing buttons, and even communicating non-verbally with humans or pets they encounter. Their design incorporates articulated limbs, stereo cameras, lidar systems, and object-recognition algorithms, making them far more adaptable than anything Amazon has built before. Importantly, Amazon is not alone in this race—other tech companies like Boston Dynamics, Agility Robotics, and Tesla are also developing humanoid platforms, but Amazon’s edge lies in its massive logistics infrastructure, access to real-world delivery data, and ability to integrate robotics with its existing fleet and operations. The move from wheels to legs isn’t just a technical upgrade—it’s a strategic response to the shortcomings of Scout and similar ground-based bots, which often fail when faced with stairs, curbs, or even simple gate latches. Humanoid robots can mimic human motion, allowing them to traverse the same environments human couriers do, without needing customized infrastructure. Moreover, coupling humanoid bots with electric vans enhances sustainability and productivity: the van acts as both a mobile hub and charging station, while the robot handles the physically demanding last steps of the delivery. However, this path is not without significant challenges. The real world is messy and unpredictable, and no amount of lab testing can account for every situation a robot might face—from an unleashed dog to a kid riding a tricycle in the path, or an icy, uneven driveway. There are also pressing concerns around privacy, safety, and public acceptance. Will customers be comfortable with a bipedal robot walking up to their homes, perhaps even being captured on their doorbell cameras? What about liabilities if a robot damages property or causes injury? These questions will require careful regulatory consideration and transparent public communication. Yet Amazon seems confident, betting that the long-term benefits—reduced labor costs, faster deliveries, increased efficiency, and alignment with its 2040 net-zero emissions pledge—will justify the investment and innovation. In fact, Amazon’s efforts go beyond hardware; they include extensive AI training environments where robots are subjected to thousands of virtual delivery scenarios before being tested in the real world. They’re taught to recognize house numbers, avoid flowerpots, and even react to dynamic changes in lighting or weather. The company is methodically training these robots to handle residential diversity—from urban brownstones to suburban cul-de-sacs—making them more adaptable than the rigid sidewalk-only Scout. In many ways, this pivot to humanoids marks a critical inflection point in robotics history. It signals a shift from novelty and experimentation to utility and integration. While Scout was symbolic of robotic ambition, the new humanoid bots represent robotic maturity—a culmination of years of AI, sensor, and mechanical advancements working in harmony toward a commercial goal. The very concept of robots jogging from vans to porches, once the stuff of sci-fi, is inching closer to daily reality, driven by the pressures of modern commerce and the relentless drive of a company like Amazon. If successful, these robots could not only change how deliveries are made, but also influence urban design, infrastructure policy, and the future of human-robot interaction at scale. Neighborhoods may soon need designated robot drop zones or path access points, and consumers might adapt to robot-friendly packaging or smart lockers. In the longer term, this technology could extend beyond delivery—servicing homes, assisting the elderly, or performing emergency tasks. For now, though, Amazon’s focus is clear: to master the last mile through machines that don’t just roll—they walk, run, and jog into a future where automation and human convenience meet at the front door.

Conclusion

Amazon’s Scout project illustrated the complexity of automating sidewalk delivery. Humanoid robots offer a compelling next step—closer to human capability with step negotiation and object manipulation. The success of real-world trials remains to be seen, but these developments could revolutionize last-mile delivery. Regulatory, ethical, and social acceptance will play crucial roles in whether customers ever "see" robots jogging up to their doors.

Q&A Section

Q1: What was Amazon Scout and why was it discontinued?

Ans: Scout was a six-wheeled sidewalk delivery robot introduced in January 2019 across U.S. test markets. It was halted in October 2022 and discontinued in early 2023, due to high operational costs, regulatory constraints, and technological limitations in climbing and doorstep navigation

Q2: What challenges did Scout face on sidewalks?

Ans: Scout struggled with unpredictable sidewalk environments—uneven terrain, obstacles like garden tools and toys, steps, privacy concerns, and frequent need for human intervention in unstructured settings .

Q3: What are these new humanoid robots Amazon is testing?

Ans: Amazon is developing electric-powered, two-legged, two-armed humanoid robots designed to deploy from Rivian vans, move independently to doorsteps, deliver parcels, and return—all trained in a “humanoid park” testing environment .

Q4: What makes humanoid robots a better solution than wheeled ones?

Ans: Humanoids can perform complex tasks—opening gates, ascending steps, balancing on uneven surfaces, manipulating objects—that wheeled robots cannot. This capability reduces need for human oversight and increases delivery accuracy.

Q5: What are the hurdles for humanoid delivery bots?

Ans: Technological challenges (dexterous manipulation, balance, vision), regulatory factors (liability, privacy, sidewalk/path access), and public acceptance in shared spaces are major hurdles before these bots can be widely deployed .