Robot.com unveils legged delivery robot to push beyond sidewalk limits

Felipe Chavez's R-dog pairs FieldAI autonomy with delivery and mobile ads, with commercial deployments planned for 2027.

By · Published

Primary source: PR Newswire

Why it matters

R-dog tests whether legged mobility and advertising revenue can extend delivery robots beyond sidewalks without breaking the economics that wheeled fleets have spent years improving.

a legged delivery robot, 'R-dog' (claymation still — hand-modelled plasticine scene with visible fingerprints)

Robot.com founder Felipe Chavez unveiled R-dog on Tuesday, a four-legged concept robot designed to carry deliveries across stairs, curbs and uneven ground, then earn additional revenue as a mobile advertising screen. Robot.com plans to move the concept into commercial deployments in 2027 through an expanded autonomy partnership with FieldAI, according to the August 11th announcement. (prnewswire.com)

The machine extends a problem Chavez has worked on since moving from Colombia to Berkeley and building Kiwi's first prototype inside UC Berkeley's MLK Student Union. Kiwi, later renamed Robot.com, initially divided deliveries into segments so small wheeled robots could handle the final stretch. R-dog revisits that architecture with a machine intended to complete the part that still requires a person when a route includes steps, door thresholds or terrain a sidewalk robot cannot cross. (news.berkeley.edu)

"A decade of building delivery robots has taught us that most robots hit a wall in the final 100 meters, leaving a person to finish the job by hand," Chavez said in the announcement. R-dog's compartment is designed to carry several food orders or e-commerce parcels, while an unattended drop-off system would release the items before the robot takes another assignment. Robot.com has not attached payload weight, battery life, range, speed or pricing to the concept, figures that will determine whether four legs can beat cheaper wheeled machines on delivery economics. (prnewswire.com)

A legged robot enters a scaled wheeled market

R-dog's mobility claim gives Robot.com a distinct pitch, though it enters a category where rivals already operate much larger delivery networks. Starship Technologies says it has completed 9 million deliveries with more than 2,700 robots, while its six-wheel design can climb and descend curbs. Serve Robotics reported more than 2,000 robots deployed across the United States by March 2026. Coco Robotics says its delivery network serves more than 3,000 merchants and restaurants. Those competitors are concentrating on fleet density, autonomy and cost per mile with wheeled hardware. (starship.xyz)

Robot.com says it has more than 500 robots deployed across the United States, Canada, Dubai and the wider Middle East and North Africa region, with more than 2.5 million tasks completed. Its task count spans a broader portfolio that includes campus delivery, warehouse logistics, advertising and other robotic work, so it cannot be compared directly with a rival's completed-delivery figure. Robot.com's operating history still gives Chavez something many robotics founders lack: years of dealing with remote assistance, merchant loading, maintenance and the mundane field failures that determine whether autonomous delivery works outside a demo. (prnewswire.com)

The competitive case for R-dog therefore rests on route completion. Starship's robots can manage curbs and snowy conditions, but Robot.com is targeting stairs and mixed indoor-outdoor paths that remain outside the normal operating envelope of a sidewalk rover. Hospitals, universities, apartment buildings and industrial campuses offer controlled starting points because the routes repeat and operators can concentrate robots within a defined area. Dense residential delivery presents a harder test, with more varied entrances and longer distances between orders. (starship.xyz)

Advertising is part of the utilization strategy

Robot.com is also trying to make R-dog useful when it is not carrying an order. Two video displays connect the robot to R-ads, Robot.com's self-serve advertising system, allowing brands to remotely manage campaigns while the machine moves through campuses, streets and event sites. The product page also promises voice interactions, campaign analytics and a sensor that makes the robot wag its tail when someone pets its head. (prnewswire.com)

Mobile advertising alone does not separate R-dog from the field. Starship already sells robot wrapping, digital messaging and product sampling around its fleet. Robot.com's wager is that a more expressive legged machine can draw people toward dual screens while reaching places a wheeled advertising robot cannot. If customers use the same hardware for deliveries during meal peaks and campaigns during events or slower periods, advertising could increase utilization and help cover the higher cost and energy demands associated with a quadruped. That outcome remains a business-model test for the 2027 deployments. (starship.xyz)

FieldAI supplies a brain it can sell elsewhere

R-dog will use Robot.com's REMI software for dispatch, teleoperation, diagnostics and fleet management. Robot.com says REMI already coordinates its delivery, cargo and humanoid machines, allowing Chavez's team to introduce another hardware form without rebuilding the operating layer. Robot.com's website says one remote supervisor currently oversees five to 10 autonomous robots and intervenes for edge cases. (prnewswire.com)

FieldAI will provide the production navigation stack. Founder Ali Agha previously worked as a robotics technologist and group leader at NASA's Jet Propulsion Laboratory after completing a PhD at Texas A&M and postdoctoral research at MIT. His Field Foundation Models are designed to reason about physics, uncertainty and risk while navigating without prebuilt maps, GPS or fixed routes. (www-robotics.jpl.nasa.gov)

That software partnership lets Robot.com focus on fleet operations, customer workflows and hardware integration instead of funding an entire autonomy research program internally. FieldAI, which announced $405 million across two funding rounds in August 2025, is pursuing a horizontal strategy: one autonomy system running across different manufacturers and robot types. RuntimeWire reported in June that FieldAI said it had crossed $100 million in revenue and contracts. (fieldai.com)

FieldAI is already applying the same approach to Boston Dynamics' Spot through a construction partnership announced in March. The arrangement gives Robot.com access to a heavily funded autonomy supplier with quadruped experience, while FieldAI remains free to improve its models through deployments on competing hardware. Robot.com's defensible layer will have to come from REMI, its customer relationships and its ability to keep a mixed fleet working reliably. (fieldai.com)

Chavez has been broadening that fleet quickly. In June, RuntimeWire examined how Robot.com was turning its delivery base toward workplace humanoids with R-noid. R-dog follows the same playbook: reuse Robot.com's operating infrastructure, bring in specialized autonomy where necessary and seek multiple revenue streams for each machine.

The 2027 pilots will show whether that platform strategy can overcome the trade-offs of legged hardware. R-dog has a clear job that wheeled rivals struggle to perform. Robot.com still has to prove that completing the final staircase is valuable enough to pay for the machine that climbs it.

Reader comments

Conversation for this story loads after sign-in.