Despite good intentions to make changes that will help the environment, companies still send items to landfills that could be recycled: the world produces almost twice as much plastic waste as it did 20 years ago, but only 9 percent of it is properly recycled, according to a report from the Organisation for Economic Cooperation and Development (OECD).
Glacier, a company that uses AI-powered robots to help companies handle the tedious and tedious task of sorting trash more efficiently, hopes to change that. Founded by former Bain consultant Rebecca Hu and former Meta software engineer Areeb Malik, the San Francisco-based company says its robots can sift through 45 items per minute (containing more than 30 different materials) to find recyclable items that human detection equipment might miss.
“Today, when you visit a facility that uses the Glacier robot, you see a conveyor belt, which is typical of these facilities. This robot has two parts. The first is an AI vision system, which is literally a camera that takes pictures of every item that passes underneath it, 24/7,” Hu says. “And then the robot has an arm that knows where to pick up an item, and it knows what the item is, so it knows where to actually sort it. And that allows us to automate the sorting process with great precision.”
That's where Glacier's computer vision technology and sorting robots come in handy: the company's analytics advise customers of the number and types of items that are missing. Glacier estimates that each robot could prevent more than 10 million items from ending up in landfills annually.
“We usually recommend that our clients start with what's called a last chance line, or a residual line,” says Foo. “In other words, after all the sorting is done, there's usually one conveyor belt, and it's the very end of the facility. Everything on that belt is supposed to be trash, but right now it's all going to the landfill because the sorting process is so difficult. You can imagine that there's a lot of really good stuff on that line as well.”
Public awareness and behavior regarding recycling continues to improve, but there is still plenty of room for improvement. According to the Environmental Protection Agency, the current recycling rate in the United States is about 32%. The agency's goal is to reach at least 50% by the end of the decade.
In March, Glacier garnered a key vote of support from Amazon, which invested in the company as part of its Climate Pledge Fund, also giving Glacier the honor of being the second female-led climate tech company backed by Amazon, which previously pledged to invest $53 million in women-led climate tech ventures.
For Hu, being part of the solution felt personal.
“I often say that this is the first time in my life that an idea is so compelling that I'm willing to take the risk of such an early-stage startup,” she says.
Hu, a first-generation American and the daughter of Chinese immigrants who grew up in the Chicago suburbs, said her family has always been resourceful when it comes to finding different ways to reuse materials.
“One of the really interesting aspects of growing up like that was there was always this mantra in the house: reduce, reuse, recycle, waste nothing. Yogurt containers were always used as storage containers for something else,” she says. “More broadly, I had this sense of being in a very consumerist society where people just threw stuff away. That was kind of a cognitive dissonance that I was constantly thinking about as I was growing up.”
With Amazon's backing and attention from companies eager to improve their environmental footprint, Hu says Glacier is focused on making its technology available to a wider audience and strengthening its AI capabilities. For example, Glacier is working on developing computer vision technology that can identify recycling waste across a wide range of categories, from wide cardboard boxes to more specific items like cat food cans.
“There's no shortage of items coming into these facilities, so with our vehicles constantly taking images, we have access to hundreds of millions of images of recyclable items that we can use as our training and validation sets,” she says. “The challenge now is not collecting those images, but figuring out what types of items should actually be detected. We're working very closely with our recycling customers and are starting to understand where most of the value is.”
While Hu is proud of the accomplishments the Glacier Robot has already demonstrated, he says humans remain a vital part of the process.
“It remains to be seen whether recycling facilities will ever become fully automated,” she says, “but the industry is already experiencing a severe labor shortage, a trend that is expected to continue for years to come. So if robots could take over the most tedious, dirty, and dangerous jobs in these facilities, much of the existing workforce could be upskilled and trained for maintenance, repair, surveillance, and other jobs. I think this would be a better use of precious human resources.”
