Advanced computer vision technology is transforming waste management from guesswork to precision, as new data reveals both progress and untapped potential in global material recovery systems.
The waste management industry is getting a serious upgrade. Greyparrot, a global leader in AI waste analytics, has just dropped its comprehensive 2025 analysis, “Recycling in 2025: Insights from 52 billion waste objects,” giving the sector unprecedented visibility into what’s really happening with recycling across more than 20 countries.
Massive scale, real-world data
Throughout 2025, Greyparrot’s network of over 200 Analyzer units observed 477 billion waste items and analysed 52 billion unique waste objects across more than 65 material recovery facilities (MRFs). The Analyzers detected 1.24 million tonnes of recyclables being processed, showing just how vital waste managers are in keeping valuable materials out of landfill and incineration and in circulation.
This dataset offers one of the most detailed looks at how materials are processed in real-world conditions, giving facilities, brands, and regulators the hard data they need to improve material recovery and strengthen circularity across global waste systems.
What the numbers are telling us
The 2025 findings paint an interesting picture. Overall, 53% of all residue material was still recyclable—meaning there’s plenty of room for improvement, but also showing that MRFs are already capturing a lot.
Paper and cardboard are still the champions of the residue stream, making up 28.9% of what’s getting left behind. Recyclable plastic accounted for 19.1% of residue material, suggesting that plastic recyclers are getting better at optimising their operations.
Here’s something worth noting: metal in residue streams jumped to 4.3% (up from 2.5% in 2024), largely due to growing e-waste volumes and safety concerns around batteries. On the bright side, glass recovery stayed strong, with only 1% showing up in residue.
Design decisions matter
Analysis using Greyparrot’s Deepnest.ai revealed something striking: removing label sleeves can make PET bottles three times more likely to be recovered. It’s a clear reminder that design choices have real consequences for recycling outcomes.
Mikela Druckman, CEO of Greyparrot, said: “Our analysis leverages Greyparrot’s global footprint to provide a detailed view of how materials are actually processed. By unlocking this waste intelligence, it offers real-world data on recycling progress, identifies where recoverable materials are lost and reveals opportunities for waste managers, brands and policymakers to improve recovery and drive systemic change.”
Turning waste into wisdom
The technology’s potential goes well beyond just running facilities more efficiently. Ambarish Mitra, Co-founder of Greyparrot, put it this way: “With AI computer vision, waste is rapidly evolving from a blind spot to a high-resolution data layer that reveals how the world really consumes and discards. We can now answer questions we have been asking for decades about consumption, recycling, pollution, design and resource management. At Greyparrot, we are unlocking this waste intelligence to amplify the voice and influence of waste management to accelerate the transition to a circular economy.”
The bottom line? While recovery rates for key materials are looking good, there are still clear opportunities to optimise how we handle fibre, plastic, and metals, and to push for design improvements that make packaging more recyclable. As AI-powered waste intelligence becomes standard across the industry, waste managers are gaining the tools they need to turn recycling from an operational headache into a strategic win for the circular economy.