Serve Robotics launches new business to become the AWS for robots
Serve Robotics announced alonsgside its third-quarter earnings that it will launch a new business line called Physical AI Services to offer customers its connectivity and data tools.
Serve Robotics, the company behind self-driving sidewalk delivery robots, is diversifying its offerings to become a supplier of technology to other robot developers. The company, which was spun off from Uber's acquisition of Postmates and went public in 2024, announced on Thursday the launch of a new business line called Physical AI Services.
This new division will provide customers with connectivity and data tools. One of Serve Robotics' customers is Bedrock Robotics, a startup that is developing self-driving systems for construction equipment. In a similar vein, the company is collaborating with T-Mobile to bring AI computing power closer to robots. According to Serve Robotics' CEO and co-founder Ali Kashani, the expansion of the company's sidewalk delivery business is shifting away from Uber and towards other partners, including DoorDash and Wonder.
Serve Robotics already produces two types of robots - a charming, expressive sidewalk delivery robot and one for use in hospitals, acquired through their acquisition of Diligent Robotics. Kashani emphasized that these businesses will continue to exist, but there is potential to sell the technology developed for their own robots to other companies that are developing robots for farms, warehouses, and other uncontrolled environments.
This expansion is part of Serve Robotics' strategy to provide technology to other companies building robots, similar to Amazon's AWS. The Physical AI Services business line will offer three products - Serve Link, a data curator tool, and a real-world, foundation-model data catalog. By providing these tools, Serve Robotics aims to help other companies build and improve their own robot applications.
Written by urgent.news from TechCrunch's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.