Coco Robotics is taking its next big leap into the future of automation. The Los Angeles-based startup, known for its fleet of last-mile delivery robots, has announced the creation of the Coco Robotics AI research lab, a new initiative focused on advancing physical AI using real-world delivery data.
The lab will be led by UCLA professor Bolei Zhou, a globally recognized researcher in computer vision, robotics, and reinforcement learning. Zhou will also join Coco Robotics as chief AI scientist, signaling the startup’s push to accelerate its shift toward full autonomy.
“We have millions of miles of data collected in complex urban environments,” said Zach Rash, co-founder and CEO of Coco Robotics. “That data is crucial for building reliable real-world AI systems — and now, with the new Coco Robotics AI research lab, we can finally scale research to the next level.”
Driving Physical AI Forward
Since launching in 2020, the Los Angeles-based startup has relied on human teleoperators to help its delivery bots navigate obstacles. However, with five years of operations and extensive real-world data, the company is ready to evolve. The new AI research lab aims to enhance robot autonomy by developing models trained on this vast dataset.
Zhou’s previous work on micromobility and robot navigation made him the ideal fit for the role. Coco Robotics had already been collaborating with him through UCLA — even donating one of its robots to the university’s research lab.
“Bolei Zhou is one of the leading researchers globally on robot navigation and reinforcement learning,” said Rash. “He’s already attracting top researchers to join the Coco Robotics AI research lab and accelerate innovation.”
Collaboration and Real-World Impact
The Coco Robotics AI research lab will operate independently from the company’s ongoing collaboration with OpenAI, which allows it to use OpenAI’s models while providing valuable robotic data for research.
Currently, the startup plans to keep the lab’s research proprietary, using insights to improve its automation systems and reduce delivery costs. Rash added that findings may also be shared with city partners to help improve infrastructure that affects delivery efficiency.
“Success for this lab means offering a higher-quality service at a much lower price,” Rash explained. “That’s how we’ll drive growth and make autonomous delivery accessible to everyone.”
A Step Toward Smarter Cities
By combining academic expertise with years of operational data, the Coco Robotics AI research lab represents a significant step toward making physical AI more efficient and accessible. For Coco Robotics, it’s not just about smarter robots—it’s about creating smarter, more connected cities.