The world has a trash problem. The amount of stuff we throw away is expected to nearly double, to 3.8 billion metric tons, by 2050. Reducing what we use would go a long way to addressing the issue, but let’s face it, we’re not very good at buying less either.
That leaves recycling, which has its own problems. People routinely try to recycle dirty yogurt cups or toss plastic in the aluminum bin. It all makes recycling more expensive because, ultimately, someone has to manually pick out the unwanted stuff.
In response, several companies have been building automated systems to sort recyclables, including Glacier, a 6-year-old company that has developed inexpensive robotic arms controlled by computer vision to identify over 30 different types of materials.
The startup has deployed its robots in San Francisco, Los Angeles, Chicago, Detroit, Phoenix, and now Seattle.
As Glacier looks to expand its robot fleet to more municipalities, it recently raised a $16 million Series A, the company exclusively told TechCrunch.
The round was led by Ecosystem Integrity Fund with participation from AlleyCorp, Alumni Ventures, Amazon Climate Pledge Fund, Cox Exponential, Elysium, New Enterprise Associates, One Small Planet, Overlap Holdings, Overture, VSC Ventures, and Working Capital Fund.
Materials recovery facilities — or MRFs, as sorting facilities are called — are getting squeezed on both ends, Rebecca Hu-Thrams, Glacier’s co-founder and CEO, told TechCrunch. Governments want more waste to be recycled, but MRFs are having a hard time finding enough people to staff the sorting line.
Industry-wide, turnover is extremely high. A typical MRF will have to hire five times per year for a single sorting position. The job is so undesirable that one MRF operator told Hu-Thrams that, even though his wages were higher, he was concerned about losing workers to a new warehouse set to open nearby.
“Would you rather stand at a conveyor belt and sort through people’s trash, or would you rather be lifting boxes in an air-conditioned warehouse?” Hu-Thrams said. “That kind of underscores the dilemma that a lot of our customers are facing.”
Glacier offers its robots to customers as outright purchases or on a lease-to-own model. It encourages MRFs to make repairs they feel comfortable with, supplying them with training and spare parts. For those that would rather not, the startup offers maintenance packages.
Glacier is also offering a data product, in which MRFs and other stakeholders like consumer products companies and government agencies can pay for access to insights about the waste stream. For an MRF, that might mean identifying where on a line it’s losing valuable aluminum cans to the landfill. For a company or regulator, it might mean auditing the waste stream to determine whether packaging that’s designed to be recycled is actually getting recycled.
With enough robots, recycling rates should improve, if only because robots are faster and better at distinguishing between recyclables and trash.
“Every time we send people to audit our AI systems, the people just do so much worse,” said Areeb Malik, Glacier’s CTO and second co-founder. “AI is getting really powerful, being able to distinguish beyond what people can even notice.”