Two Glacier programs at work in an MRF in Chicago. The robots can kind as much as 45 objects per minute. | Supply: Areeb Malik, co-founder of Glacier, on LinkedIn
Glacier, which makes use of synthetic intelligence and robotics to effectively kind recyclable supplies, right now introduced that it has raised $16 million in Collection A funding. The corporate additionally stated Recology’s King County Materials Restoration Facility, or MRF, in Seattle has deployed a fleet of its AI recycling robots.
San Francisco-based Glacier stated this funding caps a breakthrough interval throughout which it expanded its presence to markets together with San Francisco, Los Angeles, Chicago, Detroit, and Phoenix.
“The funding is just not solely an unimaginable testomony to this progress trajectory that we’re on, on our quest to allow round manufacturing; however we’re utilizing the funding in a few notable methods,” Rebecca Hu-Thrams, co-founder and CEO of Glacier, instructed The Robotic Report. “The primary is simply to make our {hardware}, together with our robotic and our imaginative and prescient programs, more and more extra scalable and sooner to deploy.”
Glacier has assembled robots to sort out the unpredictable nature of recycling streams. Its pc imaginative and prescient and AI expertise should determine and type over 30 various kinds of supplies, which might vary from frequent objects akin to PET plastic and aluminum cans to particular packaging sorts like toothpaste tubes, cat meals tins.
Glacier builds out knowledge flywheel
Hu-Thrams stated that Glacier has more and more been taking a look at its recycling robots and AI knowledge as two components of the identical flywheel. The info makes the robots higher and extra correct, whereas the robots collect extra knowledge with each choose they make.
“What we’re actually centered on now’s constructing out that flywheel and in addition understanding how we will harness that knowledge to unravel issues that will even be agnostic of what the robotic is instantly meant to deal with,” Thrams stated. “Our final imaginative and prescient is for Glacier to change into a software that may be an working system for these recycling amenities.”
The Ecosystem Integrity Fund (EIF) led the spherical. It additionally included participation from current traders together with New Enterprise Associates (NEA), AlleyCorp, Overture Local weather VC, Amazon Local weather Pledge Fund, and VSC Ventures.
New traders in Glacier included Elysium, Overlap Holdings, Cox Exponential, Alumni Ventures, One Small Planet, and Working Capital Fund.
Closing the round manufacturing loop
To know Glacier’s long-term objectives, you first want to know how the recycling worth chain works. It begins with placing recycling bins out on the curb. As soon as it’s there, the recyclables are picked up by a truck and brought to an MRF, Hu-Thrams stated.
“The job of that MRF is to principally to take that big pile of commingled recycling, a lot of which can also be trash, and type it into constituent commodities as successfully as doable,” she defined. “So if the sorting course of had been good, you’ll find yourself with these stunning bales of paper, cardboard, plastics, aluminum cans, and so forth that these MRFs might promote to finish market producers to show into new stuff. Now, the sortation course of, as you may in all probability think about, is extraordinarily cumbersome. It’s very costly, as a result of a lot of it’s nonetheless truly accomplished with handbook labor.”
MRFs expertise extraordinarily excessive ranges of turnover, and if most MRFs can’t discover the workforce, they’ll’t kind the waste into usable supplies. Hu-Thrams stated that Glacier is attempting to shut the loop in round manufacturing.
Ideally, supplies from recycling bins go to a facility, the place they’re damaged down into their particular person elements, repackaged, and delivered to a producing plant to be made into one thing else. Proper now, nevertheless, there’s a damaged hyperlink on this circle: the MRFs.
“Take into consideration what these MRFs are doing: They’re taking in feedstock, after which their job is to supply a sure yield at a sure high quality price, and do the entire high quality inspection you anticipate out of a typical manufacturing plant,” Hu-Thrams stated. “And but, they’ve so many fewer instruments than one thing like a plastic bottling plant.”
“They nonetheless have a heavy reliance on handbook labor,” she added. “They nonetheless have little by the use of knowledge at scale to assist them audit their processes and enhance effectivity. That’s actually the hole that Glacier is looking for to fill for them.”
Recycling poses sortation challenges
Glacier stated its robotic prices half as a lot as standard choices and requires simply 3 ft. of conveyor area. | Supply: Glacier
MRFs are coping with an extremely heterogeneous stream of supplies which can be soiled and really troublesome to kind. This has offered many challenges for firms hoping to automate the method. First, earlier than Glacier, different recycling sortation programs used off-the-shelf robots, which had been usually too costly for the MRF, famous Hu-Thrams.
“Secondly, a number of these models had been simply bodily giant and really space-consuming in an atmosphere that could be very space-constrained,” she added. “Think about a robotic that takes up the area of three to 4 sorters on a kind line and does the job of 1.”
Lastly, MRFs, like manufacturing vegetation, reside or die primarily based on their upstream,” Hu-Thrams stated. “This implies they require very dependable programs that may keep working regardless of the extremely variable materials stream.
“They’re not simply processing bottles and cans and paper, but additionally seeing issues like Christmas lights and surfboards and microwaves and automobile motors and even issues like weapons and hand grenades,” she stated. “How do you design a system that’s versatile sufficient to work successfully with that form of materials stream? That is why we determined that purpose-built was the best way to go.”
Glacier makes use of off-the-shelf elements from dependable OEMs, Hu-Thrams stated. The firm sourced these components independently to place collectively a robotic that’s compact, has a excessive uptime, and a sooner return on funding (ROI). Actually, Hu-Thrams stated plenty of Glacier’s clients have gotten their ROI in lower than a 12 months. For recycling, a three-year turnaround can be thought-about quick, she stated.
Register now so you do not miss out!
Glacier helps MRFs get essentially the most from the information it gathers
All through our dialog, Hu-Thrams emphasised Glacier’s dedication to placing its clients’ wants first. That is true in terms of prices, uptime, cost fashions, clients’ rights to service their robots, and, maybe most significantly, AI.
“We’ll go to our recycling facility clients and ask them, what are the commodities, or the forms of issues, that you just care essentially the most about?” Hu-Thrams stated. “We’ve truly used that strategy to flesh out the best way that we’re constructing our AI taxonomy.”
For instance, Glacier works to seek out methods for its clients to make use of their knowledge to enhance the general effectivity of the plant. “As we’ve grown out our database and as we’ve grown out the capabilities of AI, we’re truly seeing a lot of clients utilizing the information as a standalone product, like a typical manufacturing plant,” Hu-Thrams stated.
“We’ve truly helped quite a few clients determine and quantify as much as 1,000,000 {dollars} in incremental income they’re presently sending to the landfill yearly, after which we might help them fine-tune their plan and see real-time the place these adjustments made any impression to their total restoration price,” she continued.
Wanting ahead, Hu-Thrams stated that she hopes Glacier gained’t simply be a robotics or an AI knowledge firm, but additionally a companion to assist recycling amenities clear up top-line points. For instance, if a buyer was having issues with contamination in its paper stream, then Glacier might set up imaginative and prescient programs and robots in strategic places to assist enhance its purity price.
“We now, maybe for the primary time, can actually assist these amenities perceive in actual time, on the merchandise stage, what’s stepping into, what’s popping out of their bales, and what’s being ejected to the landfill,” stated Hu-Thrams. “We give them this restoration software to really make sure that they’re choosing in a selected manner.”