Coco Robotics, a startup specializing in last-mile delivery bots, announced Tuesday the establishment of a dedicated physical AI lab and the appointment of University of California Los Angeles (UCLA) professor Bolei Zhou as its chief AI scientist. This strategic initiative aims to accelerate the autonomy of the company's robot fleet by leveraging extensive data collected over five years of operation.
The Los Angeles-based company launched its operations in 2020, initially relying on teleoperators to guide its robots through complex urban delivery routes. Zach Rash, Coco Robotics co-founder and CEO, stated to TechCrunch that the company's long-term objective has always been to achieve full autonomy for its last-mile delivery robots, thereby reducing overall operational costs. Rash noted that the company has now accumulated sufficient data to pursue deeper automation efforts.
"We have millions of miles of data collected in the most complicated urban settings possible, and that data is incredibly important for training any sort of useful and reliable real world AI systems," Rash said, adding that the current data scale allows for accelerated research in physical AI.
Professor Zhou's appointment as chief AI scientist was described by Rash as a "no brainer," citing Zhou's research background in computer vision and robotics, particularly his focus on micromobility rather than larger-scale vehicles. Coco Robotics had previously collaborated with Zhou, and the company's co-founders, Rash and Brad Squicciarini, are UCLA alums who had donated one of their bots to the university's research lab. Rash highlighted Zhou's reputation as "one of the leading researchers in the whole world on robot navigation, reinforcement learning, and a lot of the technologies and areas of research that are highly relevant for us."
The new research lab is distinct from Coco Robotics' existing collaboration with OpenAI, where OpenAI utilizes the company's robot-collected data while Coco Robotics gains access to OpenAI's models. Coco Robotics intends to use the insights and research generated by its new physical AI lab internally to enhance its own automation and efficiency, specifically for the local models powering its robots. Rash confirmed the company has no current plans to commercialize or sell this proprietary data to other entities.
Rash indicated that the primary measure of the lab's success would be "offering a higher-quality service at an extremely low price," emphasizing the goal of making services more affordable for businesses and customers to foster growth within the ecosystem. The company also plans to share applicable research findings with cities where its robots operate, aiming to assist in identifying and resolving infrastructure challenges that impede robot movement.
