There is a garbage problem in the world. By 2050, the amount of things we throw away is expected to almost double to 3.8 billion tons. It will go a long way to reduce the things we use, but let's face it, we aren't good at buying less.

This leaves recycling, which has its own problems. People often try to recycle dirty yogurt cups or throw plastic into aluminum boxes. All of this makes recycling more expensive, because eventually someone has to manually pick out what they don't need.

In response, several companies have been building automated systems to sort recyclables, including Glacier, a six-year-old company that has developed cheap robotic weapons controlled by computer vision to identify more than 30 different types of materials.

The startup has deployed robots in San Francisco, Los Angeles, Chicago, Detroit, Phoenix and now Seattle.

The company specifically told TechCrunch that it recently raised $16 million in Series A as Glacier hopes to expand its fleet of robots to more municipalities.

This round is led by the Ecosystem Integrity Fund, from Alleycorp, Alumni Ventures, Amazon Climate Commitment Fund, Cox Index, Elysium, New Enterprise Associates, a small planet, Overlap Holdings, Overlay, VSC Ventures and working capital funds.

Rebecca Hu-Thrams, co-founder and CEO of Glacier, told TechCrunch that the material recovery facility (or MRFS called the sorting facility is called the sorting facility) are squeezed at both ends. The government hopes to recycle more waste, but it is difficult for the MRF to find enough people to equip the classification line.

The entire industry has extremely high turnover. A typical MRF must employ a single classification location five times a year. The work is so undesirable that an MRF operator told Hu-Thrams that he was worried about throwing workers to a new warehouse that would be open nearby, even if he had higher wages.

"Would you rather stand on a conveyor belt and sort out people's garbage, or would you rather lift the box in an air-conditioned warehouse?" Hu-Thrams said. “This emphasizes the dilemma faced by many clients.”

Glacier offers its robots to customers as direct purchase or lease to owned models. It encourages MRFs to do repairs that they are satisfied with, providing them with training and spare parts. For those who don't want to, the startup offers maintenance packages.

Glacier also offers data products where MRF and other stakeholders such as consumer product companies and government agencies can pay for insights into accessing waste streams. For MRF, this could mean determining the location of valuable aluminum cans lost in the landfill. For a company or regulator, this could mean reviewing the waste stream to determine whether the packaging designed to be recycled is actually being recycled.

With enough robots, recycling rates should be improved simply because robots are faster and better at distinguishing recyclables from garbage.

“Every time we send someone to review our AI systems, people do it worse,” said Areeb Malik, Glacier CTO and second co-founder. “Artificial intelligence has become very powerful and can differentiate between things people can even notice.”