In February 2025, Uber expanded its autonomous delivery footprint in the New York metropolitan area with a new robot food delivery service in Jersey City, New Jersey. The program is being piloted by California-based Coco Robotics, which announced the initiative in the city’s Heights neighborhood. This marks the second robot delivery operation tied to Uber Eats in Jersey City, following a downtown program launched nearly a year earlier in partnership with Avride.
Coco Robotics co-founder and CEO Zach Rash confirmed the company’s enthusiasm for the new location, telling NJ Advance Media on a Friday in mid-February that the team was “super-excited to be there.” The service is expected to begin making deliveries within the next few weeks, according to Rash. The exact launch date was not disclosed in the source material.
The Heights neighborhood pilot is notable for its timing. It arrives roughly 12 months after Uber Eats first teamed up with Avride for robot delivery in downtown Jersey City. That earlier program had the support of then-Mayor Steve Fulop, who described autonomous delivery as “a natural evolution of how cities can support local business” in a statement provided by Coco Robotics.
Fulop’s involvement with the robot delivery ecosystem is part of a broader political transition. He left office on a Wednesday in February 2025 after declining to seek a fourth term. The following Tuesday, he was set to begin a new role as president and CEO of the Partnership for New York City, a prominent business organization. The source material does not specify whether Fulop’s new position will have any direct bearing on robot delivery policy in New York City.
Coco Robotics, founded in 2020, has scaled its operations considerably since its inception. The company reports that it has produced 1,000 robots, all of which share the name “Coco.” These robots have collectively completed 500,000 deliveries in partnership with 3,000 merchant partners. The source material does not break down these figures by city or time period, nor does it specify how many of those deliveries occurred in Jersey City specifically.
The company’s business model appears to differ from traditional delivery services in a key respect. Rash explained that when Coco works directly with merchants, there is a fee per delivery that is “meaningfully less” than the cost of a human delivery driver. This suggests a pricing structure that could appeal to restaurants looking to reduce overhead, though the source material does not provide specific dollar amounts or comparative cost data.
However, Coco’s robots are not without limitations. Rash acknowledged that the robots will not venture up to the sixth floor of an apartment building. “Very rarely do we go into the buildings. Typically, we’ll meet you at the closest possible point,” he said. This means customers in multi-story residential buildings will likely need to come downstairs to retrieve their orders, a logistical detail that could influence user expectations.
The Jersey City pilot is part of a broader trend of autonomous delivery expansion across the United States. Philadelphia residents, for example, have been sharing sidewalks with Uber Eats delivery robots since early March 2025, according to the source material. That city joined a growing list of U.S. locations where robots have become a visible part of the food delivery landscape.
The source material also mentions autonomous vehicle delivery in the Phoenix area, where Uber Eats customers may receive a note during checkout that “autonomous vehicles may deliver your order.” In that scenario, a Waymo car arrives at the customer’s location, and the customer must bring their phone to pop open the trunk and retrieve the delivery. Customers can opt out of autonomous delivery during checkout if they prefer a human driver.
Waymo’s autonomous technology has a complicated history with New York. In June 2025, Waymo shared plans to bring its autonomous tech back to New York, after having first manually operated its vehicles there in 2021. The company began driving manually in the Big Apple in early July 2025, specifically in Manhattan and parts of downtown Brooklyn, as well as in nearby Jersey City and Hoboken. Waymo submitted a permit application with the New York City Department of Transportation to operate autonomously with a human behind the wheel, which was granted in late August 2025.
The source material does not specify whether the Waymo-Uber Eats integration in Phoenix or the Waymo operations in New York are directly connected to the Coco Robotics pilot in Jersey City. These appear to be separate initiatives under the broader Uber Eats autonomous delivery umbrella.
For customers using the Waymo-Uber Eats integration, the experience includes some familiar ride-hailing elements. Customers unlock the door, pop open the trunk, and start the ride from the Uber app. They are still asked to rate their ride at the end, but they are not asked to tip. If issues arise, riders can access human support 24/7 via the Uber app and from inside the Waymo vehicle, which has screens in the front and back that allow passengers to quickly summon customer support.
The source material does not provide similar customer support details for Coco Robotics’ delivery robots. It is unclear whether Coco offers 24/7 human support, in-app chat, or other assistance channels. The source material also does not disclose whether Coco robots have any onboard interface for customer interaction.
Why it matters for European robot service
The Jersey City pilot offers several lessons for European stakeholders watching the autonomous delivery sector. First, it demonstrates that major ride-hailing and food delivery platforms are willing to layer multiple autonomous solutions across a single metropolitan area. Jersey City now hosts both Avride robots in the downtown area and Coco robots in the Heights neighborhood, both operating under the Uber Eats umbrella. This multi-vendor approach suggests that platform operators may not commit to a single robotics partner, creating opportunities for multiple companies to coexist in the same market.
For European cities considering robot delivery pilots, the Jersey City example highlights the importance of neighborhood-level planning. The Heights is a distinct residential area with its own street layout and building stock, and Coco’s decision to pilot there — rather than expanding the existing downtown Avride program — indicates that different neighborhoods may require different robot form factors or operational strategies. European cities with varied urban morphologies, from medieval street grids to postwar housing blocks, may need to consider similar zoning or district-based approaches.
The building access limitation is particularly relevant for European markets. Many European cities have dense apartment buildings, often without elevators or with restricted ground-floor access. Coco’s stated policy of not going above certain floors and meeting customers at the closest possible point could face challenges in cities where ground-floor meeting points are not always convenient or safe. The source material does not specify what Coco means by “closest possible point,” nor does it address how the robot handles gated communities, secure building entrances, or other access barriers common in European urban settings.
The cost structure described by Rash — a per-delivery fee that is “meaningfully less” than human delivery — could be a key selling point for European restaurants facing labor shortages or high wage costs. However, the source material does not provide specifics on how Coco’s pricing compares to human delivery in Jersey City, nor does it indicate whether the fee structure varies by distance, order size, or time of day. European operators would need to see more granular data before making procurement decisions.
The scale figures — 1,000 robots produced and 500,000 deliveries completed since 2020 — offer a benchmark for European robotics companies. These numbers suggest that a delivery robot startup can achieve meaningful scale within five years, but they do not indicate profitability, unit economics, or maintenance costs. European investors and city officials should treat these figures as evidence of operational experience rather than financial viability.
The political dimension of the Jersey City pilot is also instructive. Fulop’s support for autonomous delivery, and his characterization of it as a natural evolution for cities, reflects a broader trend of municipal leaders embracing automation as a way to support local businesses. European city officials may face similar pressure to modernize delivery infrastructure, particularly in tourist-heavy or commercial districts where congestion and emissions are concerns. However, the source material does not provide data on emissions, traffic impact, or noise levels associated with Coco’s robots.
The Waymo integration in Phoenix and New York adds another layer to the autonomous delivery landscape. While Waymo’s operations involve autonomous cars rather than sidewalk robots, the fact that Uber Eats is simultaneously testing both approaches suggests that the platform sees value in a hybrid model. European cities may need to prepare for a future where delivery robots and autonomous vehicles operate in the same urban environment, potentially competing for curb space, sidewalk space, and customer attention.
The customer experience differences between robot and autonomous vehicle delivery are worth noting. With Waymo, customers interact with a full-sized vehicle, open the trunk, and retrieve their order. With Coco, customers meet a sidewalk robot at a designated point. Neither experience involves a human handoff, but the interaction patterns are quite different. European user research would be needed to determine which model resonates better with local consumers, as the source material does not include any customer satisfaction data.
What buyers and operators should know
For restaurants, property managers, and delivery operators considering robot services, the Jersey City pilot offers several practical takeaways, along with some notable gaps in public information.
First, the fee structure matters. Rash stated that when Coco works directly with merchants, the per-delivery fee is “meaningfully less” than what a human delivery would cost. This suggests that direct merchant partnerships are a key part of Coco’s business model. However, the source material does not disclose whether Uber Eats customers pay a separate fee for robot delivery, whether the merchant absorbs the cost, or how the pricing compares to standard Uber Eats delivery fees. Buyers should ask for transparent pricing models before committing to any robot delivery service.
Second, building access is a real constraint. Coco’s robots do not typically enter buildings and will not go above certain floors. For apartment dwellers, this means a trip downstairs to meet the robot. The source material does not specify how the robot communicates with the customer upon arrival, whether there is a time limit for pickup, or what happens if the customer does not show up. Operators in dense urban environments should clarify these logistics before launching service.
Third, the scale of Coco’s operations — 1,000 robots and 500,000 deliveries — indicates that the company has accumulated significant real-world experience. But the source material does not provide data on delivery success rates, robot downtime, maintenance intervals, or failure modes. Operators should request service-level data directly from Coco or other vendors before signing contracts.
Fourth, the multi-vendor landscape in Jersey City means that Uber Eats is not relying on a single robotics provider. This could be a deliberate strategy to hedge against supply chain disruptions or performance issues, or it could simply reflect the evolving nature of the market. Either way, operators should be aware that platform partnerships may shift over time, and they should build flexibility into their own contracts.
Fifth, the customer support model for Coco is not disclosed in the source material. While the Waymo integration offers 24/7 human support via the Uber app and in-vehicle screens, there is no equivalent information for Coco’s robot delivery. Operators should ask about support availability, response times, and escalation procedures before deploying robot delivery services.
Sixth, the source material does not mention insurance, liability, or regulatory compliance for Coco’s operations in Jersey City. It is unclear who is responsible if a robot is damaged, if a pedestrian is injured, or if a delivery is stolen. European operators, in particular, should be aware that insurance and liability frameworks vary by country, and they should seek legal advice before entering into robot delivery agreements.
Seventh, the timeline for the Jersey City launch is approximate. Rash said deliveries would begin “within the next few weeks,” but the source material does not provide a firm date. Operators should not assume that a robot delivery service will be available immediately, and they should plan for potential delays during the pilot phase.
Eighth, the source material does not specify whether Coco’s robots are sidewalk-based, curb-based, or a hybrid. The mention of meeting customers at the “closest possible point” suggests a sidewalk or pedestrian-oriented design, but the technical specifications — including robot dimensions, speed, battery life, and payload capacity — are not disclosed. Buyers should request full technical documentation from vendors.
Ninth, the merchant partner count of 3,000 suggests that Coco has broad commercial relationships, but it does not indicate the geographic distribution of those partners or the average order volume per merchant. Operators should ask for case studies or references from merchants in similar markets.
Tenth, the political context of the Jersey City pilot is worth noting. Fulop’s support for autonomous delivery was a factor in the program’s rollout, and his departure from office could influence future regulatory decisions in Jersey City. Operators should monitor local political developments, as changes in leadership can affect permit approvals, sidewalk access rules, and other regulatory matters.
Finally, the source material does not address data privacy, cybersecurity, or surveillance concerns related to Coco’s robots. Delivery robots often carry cameras and sensors, and the handling of that data is a legitimate concern for both customers and regulators. European operators, in particular, should be mindful of the General Data Protection Regulation (GDPR) and other privacy laws that may apply to robot operations.
In summary, the Jersey City pilot represents a significant step in the evolution of autonomous food delivery, but it also highlights how much information remains undisclosed. Buyers and operators should approach robot delivery services with a clear understanding of what is known — and what is not — before making commitments.
Sources
https://nypost.com/2025/02/13/lifestyle/uber-launches-robot-food-delivery-service-in-jersey-city/?jw_start={seek_to_second_number}#!
Published by Vigla Media OÜ (Estonia).