Black-I Robotics won the Chewy and MassRobotics’ CHAMP Challenge. | Source: MassRobotics
Black-I Robotics won the Chewy Autonomous Mobile Picking (CHAMP) Challenge. The challenge aimed to create a system that can address a persistent and technically complex limitation in warehouse automation: enabling fully autonomous robots to handle large, heavy, and non-rigid items within dense and dynamic fulfillment center environments.
The challenge was created by Chewy, a leading online source for pet products, supplies, and prescriptions, and MassRobotics, an independent robotics hub dedicated to accelerating robotics innovation. Chewy said it often handles large items that weigh more than 40 pounds and have variable shapes, surface textures, and levels of deformability, presenting a multi-layered manipulation challenge. Their irregular geometry and low structural stiffness reduce the effectiveness of conventional suction or parallel-jaw gripping techniques. At the same time, inconsistent stacking and presentation on pallets further complicate object recognition and grasp planning.
Black-I Robotics, a Massachusetts-based MassRobotics resident, won the $30,000 first-place prize for delivering a sophisticated, full-stack autonomous picking system. Its system featured a mobile base paired with a 6-DOF industrial arm, leveraging custom multi-modal end effectors engineered to handle large, deformable, and heavy SKUs.
Twelve global teams were selected to participate in the CHAMP Challenge, representing a diverse mix of early-stage startups and independent robotics engineers. Over several months, these teams engaged in close collaboration with members of the Chewy Robotics team, which delivered guidance on operational constraints, fulfillment workflows, and system-level requirements.
CHAMP challenge focuses on full integration
Beyond the manipulation task, the CHAMP Challenge demanded system-level integration. Robotic platforms needed to navigate through aisles as narrow as 20 inches, coordinate with live warehouse operations, and place picked items into shipping containers of varying dimensions, potentially with mixed-product contents.
The challenge called for embodied AI systems capable of perception-driven decision-making, robust grasp adaptation, and safe operation in collaborative settings. To support development, the Chewy Robotics team provided contestants with photos and videos of fulfillment operations, access to the Chewy robotics lab, and a comprehensive NVIDIA Omniverse simulation package, including a digital twin of the warehouse and 3D assets for a subset of Chewy’s product line.
The challenge aimed to enable teams to validate their systems. This included simulation-based prototypes or physical systems ready to interact with the real world.
Black-I’s approach integrated AI-driven perception with high-confidence object detection and pose estimation, enabling precise grasping of non-rigid items stacked on mixed pallets. The robot demonstrated full-facility navigation using fiducial markers and SLAM, dynamic obstacle avoidance for safe operation alongside warehouse associates, and seamless integration into downstream workflows via autonomous box placement.
The team’s consistent iteration, deep technical execution, and delivery of a complete mobile manipulation pipeline set their entry apart, MassRobotics and Chewy said. It met the challenge’s core demands for autonomy, adaptability, and deployability in constrained warehouse environments.
Arturas Malinauskas, chief engineer and founder of Breezey Machine Company. | Source: MassRobotics
Breezey Machine Company comes in second
Breezey Machine Company, a team of independent engineers from the Boston Area, came in second place and won $15,000. solution focused on end-of-arm tool innovation, presenting a novel, low-profile gripper capable of adapting to deformable and variably stacked items with minimal pre-alignment. By emphasizing mechanical compliance and passive alignment strategies, Breezey’s design achieved secure grasps without relying heavily on high-precision vision or complex control algorithms.
The team also demonstrated thoughtful consideration of integration, proposing a modular arm-mounted system that could be retrofitted to existing mobile platforms or used within compact cell configurations. Their submission stood out for its practicality, manufacturability, and the potential to serve as a robust subsystem within larger automation workflows.
Breezey’s ingenuity and attention to real-world constraints exemplified the kind of targeted, systems-level thinking the CHAMP Challenge aimed to foster.
Tell Us What You Think!