Recycleye - AI-driven robotic waste sorting platform

London, United Kingdom · 2019 · Recycleye Ltd — UK-based (London) (acquired by the US-based CP Group in 2024).

MRFs using Recycleye have moved toward more automated, hybrid sorting lines that reduce reliance on manual picking and allow reconfiguration to target high-value fractions. Environmentally, higher capture of recyclables reduces residual waste sent to landfill/incineration and improves recyclate quality for end markets. Socially, automation can improve onsite safety and reduce repetitive manual tasks, though it also requires workforce reskilling. Systemically, data and analytics from deployments inform operators and policymakers on material flows, enabling better contracting, procurement of recycled-content materials and more efficient regional recycling strategies.

Highlight

Combines advanced computer-vision AI and robotic manipulators to retrofit existing MRFs (material recovery facilities), increasing material recovery and reducing contamination on mixed recycling lines.

Extended description

Context/problem: Many material recovery facilities (MRFs) rely on manual sorting and legacy mechanical processes that struggle with complex, contaminated and variable mixed waste streams. This leads to low capture rates for valuable materials, high levels of contamination, worker safety risks, and increased landfill or incineration of recoverable resources.

How the solution works: Recycleye deploys high-resolution cameras and sensors above conveyor lines to capture continuous imagery of waste flows. Proprietary computer-vision models detect, classify and track items in real time. The system generates pick targets and control signals for robotic pick-and-place units (using FANUC 6-axis robots) or downstream actuators, enabling selective extraction of specific materials (plastics by polymer type, metals, paper, etc.). The solution is designed to retrofit into existing MRF infrastructure without a full line rebuild.

Technologies/processes involved: deep-learning image recognition, object tracking, edge computing for low-latency inference, cloud-enabled model training and analytics, and robotic manipulators or pneumatic/ejection systems for physical separation. Data pipelines collect labelled images to iteratively improve models and create material-mix analytics for operators.

Operational logic: The platform provides continuous recommendations, automated picking and operator dashboards showing stream composition and machine performance. It can prioritise valuable fractions, adapt to changing feed composition, and integrate with operators' control systems for hybrid human–robot workflows.

Stakeholders: waste-management companies and MRF operators, recyclers and commodity buyers, local authorities, equipment integrators and investors. The company also collaborates with robotics and controls suppliers (notably FANUC).

Systemic implications: By increasing recovery and reducing contamination, the solution supports circular-economy outcomes (more recyclate supply, higher-quality secondary materials), reduces landfill/incineration pressure, improves workplace safety and enables data-driven optimisation of regional waste systems. Scalability of AI models allows deployment across diverse geographies and feedstocks.

Demonstrated metrics

Real impact

MRFs using Recycleye have moved toward more automated, hybrid sorting lines that reduce reliance on manual picking and allow reconfiguration to target high-value fractions. Environmentally, higher capture of recyclables reduces residual waste sent to landfill/incineration and improves recyclate quality for end markets. Socially, automation can improve onsite safety and reduce repetitive manual tasks, though it also requires workforce reskilling. Systemically, data and analytics from deployments inform operators and policymakers on material flows, enabling better contracting, procurement of recycled-content materials and more efficient regional recycling strategies.

Implementation location & date

Founded in London in 2019. Pilots and commercial deployments primarily in the UK and parts of Europe from the late 2010s into the 2020s, used in operational MRFs as retrofit solutions (timeline and site list vary by contract). Acquired by CP Group (US) in 2024.

Sources