Inside AMP Robotics’ 120 picks-per-minute robots and Greyparrot’s 477 billion detections — proving AI’s measurable impact on recycling profitability and material recovery
In 2025, a step change occurred in the recycling industry — not in policy or design, but in operational reality. AI systems processed 52 billion unique waste items across 65 material recovery facilities (MRFs). Robots picked plastic bottles at 120 per minute, contamination dropped by 85%, and specific facilities reported 10% recovery rate improvements within months of AI deployment. AMP Robotics’ cumulative platform has now identified 150 billion items and guided the sortation of more than 2.5 million tons of recyclables, with more than 400 AI systems deployed across North America, Asia, and Europe. This case study compiles what those numbers actually mean — and the lessons every recycling operator, brand owner, and material innovator can learn.
Featured throughout is the broader context of how AI platforms on the materials design side — like Simreka and Simreka’s MatIQ – the AI Co-Pilot for Material Innovation — complete the loop between better sorting and better products made from recyclate.
Case 1: AMP Robotics at Waste Connections
Waste Connections — one of North America’s largest waste management companies — has become the largest single operator of AI-guided robotics in the industry, with 50 AMP systems booked or deployed on plastic, fiber, and residue lines. Key outcomes:
- Pick rate: Up to 120 bottles per minute per robot (vs ~40 picks/minute for humans) — a 200% throughput gain.
- Accuracy: 99% sorting accuracy across common recyclables.
- Cost: 30–50% lower operating cost than traditional manual MRF.
- Recovery: >90% recovery of targeted recyclables.
Case 2: AMP + Evergreen Clyde Facility
At Evergreen’s Clyde, Ohio facility, AMP’s robots removed up to 90% of contamination on average across different lines. The Ellen MacArthur Foundation analysis found AI-driven sorting reduces contamination rates by up to 85% — translating directly to higher-quality, higher-priced recycled plastic bales. The Clyde site ships food-contact-grade rPET flake to beverage bottle converters, with bale purity averaging 98.4% through 2025.
Case 3: AMP + SPSA Virginia — Landfill Life Extended 35 Years
In late 2024, AMP announced a 20-year contract with the Southeastern Public Service Authority (SPSA) of Virginia to process solid waste for eight communities serving 1.2 million residents. The agreement guarantees diversion of 50% of waste from landfill, with the combined AI sorting and organics management capable of processing 540,000 tons annually — extending the local landfill’s life by another 35 years.
Case 4: AMP ONE at Waste Connections Commerce City (2026)
Commissioned in early 2026, the Commerce City, Colorado facility is the first greenfield MRF in North America designed from the ground up around AMP ONE — AMP’s integrated end-to-end AI sortation platform. Initial throughput: up to 62,000 tons of single-stream recycling annually. Early 2026 operating data show 99% recoverable-material identification, with residue rates under 6% — less than half the industry average.
Case 5: Greyparrot Analytics Across 65 Global MRFs
Greyparrot’s Analyzer units, positioned above conveyor belts, capture imagery and generate real-time material composition data. Their 2025 “Unwrapped” report revealed the scale:
- 477 billion bounding box detections across the fleet.
- 52 billion unique waste objects analyzed.
- 1.24 million tonnes of recyclable material processed across 65 recovery facilities worldwide.
Case 6: Murphy Road Recycling — Facility-Wide AI Monitoring
Murphy Road Recycling’s “All-American MRF” in Connecticut became the first US operation to deploy facility-wide AI monitoring with 15 Greyparrot Analyzer units. Waste intelligence is embedded in day-to-day operations: every line, every shift, every commodity is continuously characterized. This granular visibility enabled Murphy Road to redirect previously-residue PP thermoforms to a new dedicated recovery stream, adding estimated $1.8M/year in bale revenue.
Case 7: KSI Recycling — 10% Recovery Uplift from Analytics Alone
KSI Recycling used Greyparrot Analyzer data to optimize its cleaning schedule, boosting recovery by about 10%, reducing loss of valuable plastics, and maintaining residue composition within regulatory thresholds. This demonstrates that AI-driven data analytics — even without adding robots — can unlock significant efficiency gains in existing plants.
Case 8: Dutch MRF — 10% Recovery Gain with Analyzer Data
One of the Netherlands’ leading recycling operators increased recovery by 10% using Greyparrot Analyzer data, proving consistent results across regional waste streams and operator practices.
Case 9: UK MRF — 93% of “Residue” Was Actually Recoverable
A UK-based MRF using Greyparrot analytics discovered that only 7% of its residue line was genuinely non-recyclable — the remaining 93% was recoverable. This finding alone redefined investment priorities and process tuning.
Summary of AI Recycling Impact Metrics
| Case Study | AI System | Key Outcome | Business Impact |
|---|---|---|---|
| Waste Connections | AMP Neuron + robots (50 deployed) | 120 picks/min, 99% accuracy | 30–50% lower OPEX |
| Evergreen Clyde | AMP Robotics | 90% contamination removal | Higher bale prices (98.4% purity) |
| SPSA Virginia | AMP Neuron + organics system | 50% landfill diversion | 35-year landfill life extension |
| Commerce City (AMP ONE) | AMP ONE greenfield | 62,000 t/yr, <6% residue | Benchmark greenfield MRF |
| Greyparrot Fleet (2025) | Analyzer + cloud analytics | 1.24M tonnes / 52B items analyzed | Industry-wide insight |
| Murphy Road | 15 Greyparrot Analyzers | Facility-wide AI monitoring | $1.8M/yr incremental revenue |
| KSI Recycling | Greyparrot Analyzer | 10% recovery uplift | Regulatory compliance + margin |
| Dutch MRF | Greyparrot Analyzer | 10% recovery gain | Operational optimization |
| UK MRF | Greyparrot Analyzer | 93% residue recoverable | Reinvestment in recovery |
Key Lessons for Recycling Operators
Lesson 1: Data alone is transformational — robots amplify it
The KSI and Dutch MRF cases show that simple analytics — no robots added — can unlock 10% recovery gains by informing better operating decisions.
Lesson 2: Contamination reduction pays twice
Higher purity means both higher-priced bales and fewer rejections by end-market customers. The 85% contamination reduction becomes a margin lever.
Lesson 3: Design-side AI closes the loop
Recovery is capped by what enters the MRF. Upstream platforms like Simreka’s Virtual Experiment Platform and Simreka’s AI-Powered Formulation Generator help brand owners design products that MRFs can actually recover at high value.
Lesson 4: Recyclate quality drives reformulation economics
With 99% clean bales flowing out of AI-sorted MRFs, formulators can reliably incorporate 30–50%+ recycled content without quality drift — especially when paired with Simreka’s formulation platforms that account for recyclate variability. Simreka’s Databank tracks recyclate-grade properties from the most-active AI-sorted MRFs, letting product teams plan formulations against real supply profiles.
Lesson 5: Regulatory alignment matters
The UK Environment Agency is reviewing Greyparrot’s AI sampling methodology for regulatory approval — positioning AI as a compliance tool, not just an efficiency tool.
Case 10: CP Group + Recycleye — Vision AI Meets Air-Jet Ejection
In April 2026, CP Group acquired Recycleye, consolidating Recycleye Vision (28-class classification) and Recycleye Qualibot (33,000 picks per 10 hours) with CP Group’s materials-recovery equipment portfolio. Early integrated pilot lines in Florida and Texas reported 12% gains in residue recovery and 45% lower picker maintenance costs compared to mechanical delta robots. The deal signals how AI-first sortation is consolidating into vertically integrated vendor stacks — akin to the robotics consolidation that reshaped automotive automation in the 2010s.
Case 11: AMP ReClaim Small-Format Recovery
AMP’s ReClaim system, deployed at five pilot MRFs through 2025, targets the notoriously hard-to-recover small-format stream (<5 cm). Early results show 68% PP cap recovery and 52% small-format PET thermoform recovery — substantial uplift from the near-zero baseline most MRFs achieve today. If scaled industry-wide, small-format recovery alone could add an estimated $850M/year in global bale revenue.
Integrating Case Study Lessons Into Your Strategy
- Start with data. Deploy an Analyzer-class AI to understand your current stream before buying robots.
- Prioritize high-value streams. PET bottles, HDPE containers, and aluminum cans deliver the fastest ROI.
- Design for what you can recover. Feed MRF composition data back to brand-owner product design teams — enabled by Simreka’s integration capabilities.
- Formulate with recyclate. Use Simreka AI platforms to ensure product specs are met despite feedstock variability.
- Measure and publish. Transparent, auditable recovery rates meet growing regulatory and investor scrutiny.
- Plan for vendor consolidation. The CP/Recycleye deal previews a wave of M&A; procurement contracts should include roadmap and data-portability clauses.
Conclusion
The numbers are no longer theoretical. AI-driven recycling delivered 10% recovery uplifts, 85% contamination reductions, 90% material recovery, and 120-picks-per-minute throughput across real facilities serving millions of people in 2025. AMP’s platform hit 150 billion items, Greyparrot crossed 52 billion unique objects analyzed, and landmark greenfield facilities like Commerce City set a new benchmark for what an AI-first MRF looks like. Paired with upstream AI-driven formulation platforms like Simreka, the entire plastic circular loop is now instrumented, optimized, and measurably profitable. The recycling industry has entered its AI era — and there is no going back.
Frequently Asked Questions
Q1. How much faster can AI robots sort plastic than humans?
AMP’s AI-powered robots achieve up to 120 picks per minute per arm — roughly 200% faster than a human operator’s ~40 picks/minute.
Q2. How much contamination can AI remove?
Real deployments report up to 90% contamination removal, and Ellen MacArthur Foundation analysis found up to 85% contamination reduction across AI-sorted lines.
Q3. What is the total scale of AI recycling analytics today?
Greyparrot alone analyzed 52 billion unique waste items and 1.24 million tonnes of material across 65 facilities in 2025. AMP’s platform has cumulatively identified 150 billion items.
Q4. How long is the ROI for an AI recycling investment?
Operators report 2–4 year payback periods for AI robots and analytics, driven by 30–50% lower OPEX, higher bale prices, and improved recovery rates.
Q5. Can AI recycling work in small facilities?
Yes — compact AI sortation units and analytics-only deployments (like Greyparrot Analyzers) make smaller facilities viable without full robotic retrofits.
Q6. How does Simreka extend the value of AI recycling?
AI recycling produces high-purity, high-variability recyclate. Simreka’s AI formulation platforms let brand owners design products that maintain performance with that recyclate — closing the loop on value, not just material.
Bibliographical Sources
- Fortune. “Will recycling ever be profitable? AMP Robotics is using AI to make it so.” https://fortune.com/2025/06/26/ai-recycling-trash-amp-robotics/
- US Plastics Pact. “AMP Robotics: AI Platform for Small Format Recycling.” https://usplasticspact.org/case-study/amp-robotics/
- Greyparrot. “Recycling in 2025: What we learned from 52 billion waste items.” https://www.greyparrot.ai/resources/blog/2025-unwrapped
- Greyparrot. “3 ways to drive profitable plastic recovery with AI waste analytics.” https://www.greyparrot.ai/resources/case-studies/waste-analytics-drive-profits
- Ellen MacArthur Foundation. “Artificial intelligence for recycling: AMP Robotics.” https://www.ellenmacarthurfoundation.org/circular-examples/artificial-intelligence-for-recycling-amp-robotics
- AMP Sortation. “AI Takes Out the Trash: Largest U.S. Recycling Project (SPSA).” https://ampsortation.com/articles/largest-us-recycling-project-spsa
- Recycling Today. “How AI is Helping Improve Plastic Recycling Efficiency.” https://www.recyclingtoday.org/blogs/news/how-ai-is-helping-improve-plastic-recycling-efficiency
- Greyparrot. “Greyparrot’s 2025 highlights: Turning circular momentum into intelligent action.” https://www.greyparrot.ai/resources/blog/greyparrot-2025-highlights
- Greyparrot. “Running a profitable MRF in 2026: What AI really changes.” https://www.greyparrot.ai/resources/blog/running-a-profitable-mrf-in-2026
- Robotics & Automation News. “CP Group acquires Recycleye to expand AI-driven sorting.” https://roboticsandautomationnews.com/2026/04/15/
Close Your Loop with Simreka
AI is transforming plastic recovery — now make the recyclate work in your products. Simreka helps brand owners and formulators design products that maintain performance with real-world recycled content.
Request a Demo of Simreka’s AI Platform →
