At Packaging World’s 2025 Packaging Recycling Summit in Dallas, speakers explored how artificial intelligence is reshaping recycling for both material recovery facilities (MRFs) and brand owners. It’s no longer just about faster sorting or better automation. AI is bringing sharper intelligence to every belt and bale, helping MRFs move from guesswork to real-time decision-making that boosts recovery and reduces losses. At the same time, it’s giving brands unprecedented visibility into how their packaging performs in the real world, offering data that can inform better design choices and support stronger sustainability claims.
That potential for smarter, data-driven recovery is especially critical given the scale of today’s losses. “Twenty-five percent of what’s coming into MRFs is being lost,” said JD Ambati, founder and CEO of EverestLabs, a provider of AI and robotics systems designed to improve the sorting and recovery of recyclables. “We’re losing tens of billions of dollars of materials.”
 | Read this related article from PRS24, “Colgate, Amazon Report ‘History-Making’ Progress with AI for Recycling” |
And it’s not just plastic. He also pointed to aluminum and fiber as major value streams that are slipping through the cracks, explaining that the only way to improve recovery is to make the system smarter. “This is a data problem,” he said. “And thanks to declining compute costs, we can now capture this data in near real time and act on it.”
It’s a shift that’s already happening inside progressive MRFs. For Rumpke, AI has become a critical part of daily operations, helping pinpoint what’s coming in, what’s getting missed, and how to recover it more effectively. At its Columbus and Louisville facilities, Rumpke is using AI-powered robotic sorters equipped with tools like X-ray and laser tagging to identify and remove materials automatically. “We’re seeing through the burden depth,” Snyder said, noting how the technology is enhancing both labor efficiency and sortation quality.
Greyparrot
On the technology side, Gaspard Duthilleul, COO of AI waste analytics provider Greyparrot, described how its systems support that shift. Using computer vision, Greyparrot tracks not just material types, but also branded packaging, formats, and contamination levels across the line.
“Ten years ago, it would’ve taken four years of manual sampling to get the kind of data we now collect in a few months,” he said. In one study conducted along with Closed Loop Partners, Greyparrot recorded 40 million data points in three months across four U.S. MRFs, mapping out what was actually recoverable and what was lost.
Greyparrot’s latest product launch, Deepnest, aims to bridge the gap between brands and recycling operations by giving CPGs unprecedented visibility into how their packaging behaves in a MRF. Deepnest allows brand owners to see what types of packaging are showing up at sorting facilities, how they’re performing, and where they may be falling short. It’s a tool designed to move packaging data out of the black box and into actionable territory, supporting better design, stronger claims, and more informed EPR strategies.
That kind of transparency is a game-changer for both MRF operators and brands. “With robust label sets and vision systems, we can now tell brand owners exactly how many of their bottles—or their competitors’—are moving through our lines,” said Davis of AMP Robotics. That data not only supports EPR, but it’s also helping brands see whether their packaging is truly making it through the system.
Marcinko of WM noted that this insight is shaping how WM operates. “The data we’re getting from AI is influencing our decisions around equipment, around training, and even around how we work with brands,” he said. “It’s creating a feedback loop we didn’t have before.”
That feedback loop is essential not just for optimizing current operations but for planning future investments. At Balcones Recycling, AI insights are being used to make the business case for upgrades. “It helps us understand what we should be doing, what we’re not doing yet, and what’s actually worth recovering,” said Mariel.
Still, as Marcinko cautioned, not all MRFs are created equal. “There are over 180 facilities across the U.S., most of them mechanical, and very few have advanced AI yet,” he said. That means while the data revolution is underway, its full potential won’t be realized until adoption is more widespread, and funded.
| | Read this related article, “Live at Waste Expo: Brands to Benefit from MRFs’ AI Robotic Sorters?” |
To that end, Closed Loop Partners has stepped in to help accelerate deployment. “We invested in Greyparrot because we saw the potential of AI to reduce waste to landfill, improve material quality, and make the entire system more efficient and resilient,” said Kristin Taylor, director of catalytic capital at Closed Loop. “It’s not just about brands or MRFs, it’s about connecting all the dots across the value chain with real-world insights.”
Looking ahead, speakers were optimistic that AI would help close the longstanding gap between design and recovery. “We talk a lot about collaboration, but not enough of it happens,” Duthilleul noted. “With data, we can start having those conversations based on facts.” PW
