Playbook: adopting Recycling systems & material recovery in 90 days

In 2024, the global circularity gap reached a critical inflection point: secondary material use dropped 21% between 2018 and 2023, signalling an alarming increase in virgin material reliance even as recycling infrastructure expanded (Greyparrot, 2024). The UK, despite its ambitious Extended Producer Responsibility (EPR) scheme and £1.1 billion government recycling investment announced in 2025, still captures only 44% of recyclable packaging—leaving more than half of recoverable materials destined for landfill or incineration. Meanwhile, the US plastic recycling rate has plummeted to just 5%, the lowest among developed nations, while the EU achieves 41% for plastic packaging (EPA Infrastructure Assessment, 2024). For organisations seeking to implement or upgrade material recovery systems, the gap between policy ambition and operational reality represents both a challenge and an opportunity. This playbook provides a structured 90-day framework for deploying recycling and material recovery infrastructure, drawing on the latest advances in AI-powered sorting, optical separation technologies, and data-driven process optimisation.

Why It Matters

The business case for investing in recycling systems extends far beyond regulatory compliance. The Recycling Partnership's 2024 State of Recycling Report found that with $36.5–43.4 billion in infrastructure investment, recovery rates could reach 61% by 2030—enabling the capture of an additional 82–89 million tons of material annually. This represents a 91% increase from 2018 baselines and would surpass the EPA's 50% national recycling goal.

For UK organisations specifically, the convergence of three regulatory drivers makes action urgent. First, the EPR scheme shifts the financial burden of packaging waste management onto producers, creating direct cost incentives for recyclability. Second, the Deposit Return Scheme launching in 2027 will transform collection economics for beverage containers. Third, EU directives requiring 35% recycled content in plastic packaging by 2030 (rising to 60% by 2040) affect UK exporters and supply chains.

Beyond compliance, material recovery delivers measurable operational benefits. Organisations implementing advanced sorting systems report 30–50% reductions in labour costs through automation (AMP Robotics, 2024). The LRS Exchange MRF in Chicago, which won the 2024 NWRA Recycling Facility of the Year Award, processes 500 tons daily with AI-powered sorting—demonstrating that modern facilities can achieve both scale and precision.

Key Concepts

Material Recovery Facility (MRF) Fundamentals

A Material Recovery Facility serves as the critical node between collection and reprocessing. Modern MRFs employ three primary sorting mechanisms: mechanical separation (screens, air classifiers, magnets), optical sorting (near-infrared spectroscopy, visual cameras), and increasingly, AI-powered robotic pickers. The evolution from single-stream manual facilities to automated multi-material plants has transformed throughput economics: optical sorters now achieve speeds of up to 8 tons per hour, compared to 0.5 tons per hour for robotic units focused on precision picking (CP Group, 2024).

Capture Rate vs. Recovery Rate

Understanding the distinction between capture rate and recovery rate is essential for setting meaningful targets. Capture rate measures the percentage of recyclable material successfully collected from the waste stream. Recovery rate measures what percentage of collected material is actually recycled rather than rejected or contaminated. The Recycling Partnership reports that only 21% of recyclable material is captured in the US, but even captured material faces further losses at MRFs due to contamination and sorting errors.

Life Cycle Assessment (LCA) Integration

Effective material recovery programmes require LCA integration to ensure net environmental benefit. Not all recycling is equal: the energy and emissions associated with collection, transport, and reprocessing must be weighed against virgin material alternatives. For aluminium, the calculus is unambiguous—recycling uses 95% less energy than primary production. For mixed plastics, particularly those requiring chemical recycling, the picture is more nuanced and demands rigorous LCA protocols.

Additionality and Counterfactual Analysis

Additionality—demonstrating that recycling activities create environmental benefits beyond what would have occurred anyway—has become increasingly important for organisations making sustainability claims. The integrity of recycling credits and EPR compliance depends on robust chain-of-custody documentation and counterfactual analysis.

What's Working

AI-Powered Sorting and Computer Vision

The integration of artificial intelligence into MRF operations represents the most significant advancement in recycling technology since optical sorting. Glacier, backed by Amazon's Climate Pledge Fund, has developed robots capable of identifying 30+ material categories including challenging items like black plastics and shrink-sleeve PET (Resource Recycling, 2024). Their systems provide real-time analytics dashboards that enable continuous process optimisation.

TOMRA's AUTOSORT platform combines NIR spectroscopy, visual cameras, and the proprietary DEEP LAISER system (laser plus AI camera) to detect materials that traditional optical sorters miss. This multi-sensor fusion approach achieves purity levels that command premium prices for recovered commodities—clear PET sorted as Grade A/B rather than C/D can fetch significantly higher prices.

Modular and Distributed Infrastructure

The traditional model of centralised mega-MRFs is giving way to distributed, modular approaches. Kiverco in the UK develops recycling plants that can be operational within one day, enabling rapid deployment in response to regulatory changes or market opportunities. This modularity reduces capital risk and allows organisations to scale capacity incrementally based on actual throughput.

Data-Driven Process Optimisation

Greyparrot's AI-powered waste analytics platform has analysed over 40 billion waste objects, providing unprecedented visibility into material flows and contamination sources. Their partnership with Bollegraaf Group, announced in 2024, integrates this analytical capability with conveyor and sorting hardware—creating closed-loop systems that continuously improve performance.

Sector-Specific KPI Benchmarks

Metric	Baseline	Good	Excellent	Unit
Capture Rate	<30%	50-65%	>75%	% of recyclables collected
Contamination Rate	>25%	10-15%	<8%	% of sorted material rejected
Sorting Throughput	<15	25-40	>50	tons/hour
Material Purity (PET)	Grade C/D	Grade B	Grade A	bale quality
Downtime	>15%	5-10%	<3%	% of operational hours
Labour Cost per Ton	>£35	£20-30	<£15	£/ton processed

What's Not Working

Single-Stream Contamination

The widespread adoption of single-stream recycling, while convenient for households, has created persistent contamination problems. Wishcycling—placing non-recyclable items in recycling bins based on aspiration rather than knowledge—contaminates entire loads. The EPA estimates that contamination rates in single-stream systems average 25%, with some facilities reporting rates exceeding 40%.

Plastic Complexity and Labelling Failures

The proliferation of plastic types, multi-layer packaging, and inconsistent labelling defeats even advanced sorting systems. Full-sleeve labels obscure underlying materials from optical sorters. Black plastic, historically invisible to NIR sensors, remains problematic despite recent DEEP LAISER advances. The chasing arrows symbol on packaging often indicates material type rather than recyclability, confusing consumers and increasing contamination.

Economic Volatility in Commodity Markets

Recycled commodity prices fluctuate dramatically based on virgin material costs, export market access, and demand from manufacturers. When virgin plastic prices fall—as they did following 2020 oil market disruptions—the economic case for recycled content weakens. Organisations dependent on commodity sales for MRF viability face structural business model challenges.

Infrastructure Gaps and Collection Deserts

The Recycling Partnership found that 80% of US states lack deposit programmes, and many households have no equivalent recycling access to their trash services. The UK faces similar challenges in rural areas and multi-occupancy buildings. Advanced MRF technology cannot compensate for fundamental collection infrastructure gaps.

Key Players

Established Leaders

TOMRA (Norway/UK): The global leader in reverse vending and sensor-based sorting, TOMRA's systems process materials in over 100,000 installations worldwide. Their AUTOSORT line represents the current state-of-the-art in optical sorting technology.

Veolia (France/UK): Operating one of the UK's largest networks of MRFs, Veolia combines collection, processing, and remanufacturing capabilities across the circular value chain.

CP Group (US): Provider of complete MRF systems including the Vivid AI platform, CP Group equipped the award-winning LRS Exchange facility in Chicago.

Biffa (UK): A leading UK integrated waste management company operating multiple MRFs and investing heavily in plastics recycling infrastructure.

Emerging Startups

Greyparrot (London): AI waste analytics pioneer with £20M+ in funding, providing computer vision systems that have analysed 40+ billion waste objects. Their 2024 partnership with Bollegraaf signals growing hardware-software integration.

Sorted (London): Raised £1.65M in April 2024 for AI plus laser technology that helps pickers extract recyclables with real-time guidance and data insights for managers.

Impact Recycling (UK): Developer of BOSS (water-based density separation) technology, with €3.8M funding in 2025 and partnerships including Nestlé and LG Chem.

AMP Robotics (US): Leading AI robotics provider whose "pervasive AI" approach enables second-by-second waste characterisation across entire facilities.

Key Investors and Funders

Amazon Climate Pledge Fund: Backed Glacier's $7.7M raise, signalling strategic interest in MRF automation.

Breakthrough Energy Ventures: Bill Gates' climate fund has invested across the circular economy including advanced recycling technologies.

WRAP (UK): The government-backed Waste and Resources Action Programme provides grants, technical assistance, and research funding for recycling innovation.

UK Government: The £1.1B recycling investment announced in 2025 represents the largest public commitment to recycling infrastructure in UK history.

Examples

1. LRS Chicago MRF (US)

LRS invested $50 million in their Exchange MRF, incorporating CP Group systems with Vivid AI visual recognition. The facility processes 500 tons daily and won the 2024 NWRA Recycling Facility of the Year Award. Key innovations include pervasive camera coverage enabling real-time contamination tracking and adaptive sorting parameters. The facility demonstrates that large-scale single-stream processing can achieve Grade A bale quality with appropriate technology investment.

2. Rumpke Columbus MRF (US)

Opened in May 2024, Rumpke's Columbus facility processes 65 tons per hour using advanced automation and optical sorting. The facility serves as a regional hub, aggregating material from multiple collection programmes. Their approach emphasises modularity—infrastructure designed for future expansion and technology upgrades without fundamental redesign.

3. CSWD Vermont MRF (US)

Chittenden Solid Waste District committed $38 million to a new MRF after years of siting challenges. The project illustrates the often-underestimated importance of community engagement and permitting in recycling infrastructure development. Their extended planning process, while delaying implementation, ultimately secured community support that will enable long-term operational stability.

Action Checklist

Days 1–30: Assessment and Planning

Conduct baseline waste characterisation study using sampling protocols aligned with WRAP guidance
Map existing collection infrastructure and identify capture rate gaps
Benchmark current contamination rates against sector KPIs (target: <15%)
Engage key stakeholders including council waste teams, collection contractors, and downstream reprocessors
Evaluate technology options: optical sorting, AI robotics, mechanical systems
Develop financial model incorporating EPR fee implications and commodity price scenarios

Days 31–60: Procurement and Design

Issue RFP to qualified system integrators (TOMRA, CP Group, Machinex, BHS)
Conduct site assessments for MRF location including transport access and utility capacity
Finalise technology selection based on material mix, throughput requirements, and budget
Negotiate contracts with attention to performance guarantees, maintenance terms, and upgrade paths
Develop staff training programme covering safety, equipment operation, and quality protocols
Establish data systems for real-time monitoring and regulatory reporting

Days 61–90: Implementation and Launch

Complete equipment installation and commissioning
Execute staff training and competency certification
Conduct operational testing with controlled material inputs
Establish downstream offtake agreements for sorted commodities
Launch with defined KPI tracking and continuous improvement protocols
Schedule 30-day post-launch review with system integrator

FAQ

Q: What is the typical payback period for AI-powered sorting equipment?

A: Payback periods vary based on throughput volume and labour cost displacement, but most organisations report 2–4 year payback on AI robotics investments. Glacier and AMP Robotics systems typically reduce labour costs by 30–50%, and the continuous uptime of robotic sorters (compared to shift-based manual operations) accelerates returns. Facilities processing over 50,000 tons annually generally see faster payback than smaller operations.

Q: How should organisations prepare for the UK's Extended Producer Responsibility scheme?

A: EPR preparation requires three parallel workstreams. First, establish accurate data on packaging placed on market by material type and recyclability. Second, engage with compliance schemes that will aggregate producer obligations. Third, review packaging design to maximise recyclability and minimise EPR fees—the modulated fee structure penalises difficult-to-recycle formats. Organisations should also consider forward contracts with recyclers to secure processing capacity.

Q: What contamination rate should we target, and how can it be reduced?

A: Best-in-class MRFs achieve contamination rates below 8%, though 10–15% represents a realistic target for most operations. Reduction strategies include upstream education programmes (reducing wishcycling), positive sorting AI that identifies and ejects contaminants, and quality feedback loops that trace contamination to specific collection routes. WRAP's Recycle Now campaign materials provide evidence-based messaging for consumer education.

Q: How do we evaluate between optical sorting and AI robotics?

A: The technologies are increasingly complementary rather than competitive. Optical sorters excel at high-throughput bulk separation (up to 8 tons/hour) and should form the backbone of material separation. AI robotics add value for quality control picking, handling materials that optical systems miss, and providing granular data on material flows. Modern MRF designs typically layer both technologies.

Q: What metrics should we report to demonstrate recycling programme effectiveness?

A: Core metrics include capture rate (% of available recyclables collected), contamination rate, material purity by commodity, processing cost per ton, and carbon intensity (CO2e per ton processed). For EPR compliance, chain-of-custody documentation and evidence of actual recycling (not just collection) is increasingly required. WRAP's guidelines provide standardised reporting frameworks aligned with regulatory expectations.

Sources

EPA (2024). US Recycling Infrastructure Assessment and State Data Collection Reports. United States Environmental Protection Agency.
Greyparrot (2024). What We Learned by Detecting 40 Billion Waste Objects in 2024. Greyparrot Resources.
The Recycling Partnership (2024). 2024 State of Recycling Report. The Recycling Partnership.
Resource Recycling (2024). Why Glacier's MRF Robots Drew Millions in Funding. Resource Recycling.
AMP Robotics (2024). AMP Lays Out Vision of Next-Generation, AI-Driven MRFs. Resource Recycling Plastics.
UK Government (2025). £1.1 Billion Recycling Investment Announcement. UK Startup Blog / Government Press Release.
WRAP (2024). Guidance on Packaging Extended Producer Responsibility. Waste and Resources Action Programme.
AF&PA (2025). Paper Industry Announces 2024 US Paper Recycling Rates. American Forest & Paper Association.