TL;DR

Greyparrot has raised £20m ($27m) in Series B funding led by technology investor Omar Mir to scale its AI waste intelligence platform. Its camera systems, called Analyzers, sit above sorting belts in recycling facilities across more than 20 countries, identifying materials, products and brands in real time. Veolia, Biffa and FCC are among the operators using them.

The reported returns are unusually concrete for an AI deployment. Facilities using the real-time data report efficiency gains of 10% to 30%, with one site saying it saves more than £1.5m a year. Greyparrot says the funding will let it scale across North America and Europe and grow its AI, data science and product teams, targeting the abatement of more than a million tonnes of waste by 2030.

The measurement gap it addresses is stark. Less than 0.1% of the world’s estimated 2.3 billion tonnes of annual solid waste is currently audited, leaving operators, brands and regulators with little visibility of where materials are lost. Pew and Systemiq estimate $80bn to $120bn of plastic value disappears through linear systems each year. Across Greyparrot’s own network, the materials its Analyzers identify represent an estimated £1.87bn in recoverable value, including 17.4 billion PET bottles and 4.2 billion aluminium items — a metal recyclable using up to 95% less energy than primary production.

A second business has grown out of the first. Consumer goods companies including Unilever, L’Oréal and Kenvue use the Deepnest platform to inform packaging redesign and manage exposure to Extended Producer Responsibility fees and the EU Packaging and Packaging Waste Regulation. The same verified data feeds regulators and statutory reporting; Biffa and FCC were among the first waste firms globally to use AI for compliance reporting.

Co-founder Ambarish Mitra framed the ambition as infrastructure: waste intelligence will “do for materials what satellite data did for navigation”.

Looking Forward

The compliance angle is what makes this durable rather than cyclical. EPR schemes and the PPWR turn material composition into a reportable, chargeable quantity, which converts Greyparrot’s data from an efficiency tool into a regulatory requirement — demand that does not soften when capital markets do. For UK readers it is also a reminder that computer vision applied to a physical industrial problem remains a category where British firms compete on equal terms, unlike frontier model development.