This clawed robot learns from every bottle to cut sorting costs

The Danu Robotics recycling robot uses cameras and machine learning to spot different materials on a sorting line, then picks them off with a pincer claw. Most rival pickers, such as those from Glacier and RecycleEye, use suction instead, sucking items up like a vacuum.

A claw can grab objects that suction struggles with, though the bigger difference is in the software. Danu uses AI to keep improving how HERO identifies and handles items, so the system gets better the longer it runs. In plain terms, every item it handles adds to what it knows.

Ma also designed HERO as a team of several arms working together, rather than one big robot arm. She has said the idea is for it to behave like a group of human pickers, only faster, more accurate and much cheaper to run.

Sorting is where recycling often breaks down. The world produces more than three billion tonnes of waste a year, and less than 10% of it gets recycled, Ma has said. Most facilities still depend on people sorting by hand, a hazardous job with high staff turnover and rising costs.

Poor sorting has a price. Contamination rates can reach 50%, which can bring fines for waste companies. Contaminated loads also lose value, need a second costly round of sorting, or get sent to landfill or burned.

Robots are not new to the industry, but Ma says many operators that tried them later switched them off because they didn’t do the job well. That track record is the doubt Danu has to overcome. Ma admits she met plenty of skepticism early on, so she built a working prototype quickly and used real-world trials to gather the performance data that buyers and investors wanted to see.

The pitch to recycling centers is simple: they earn money from the materials they manage to recover, so a better sorter pays off straight away. Ma estimates a working site earns about $485,000 in extra revenue from having HERO on the line, which she says is more than double the robot’s cost.

Early demand looks real, if still small. Danu has $500,000 in signed contracts, letters of interest from two large customers and more than 200 potential buyers in its sales pipeline. That pipeline points at a sorting market worth about $20 billion in Europe and North America combined.

The robots are already shipping, and Danu has run trials with Glasgow City Council and works with both Glasgow and Dundee city councils. Before that, the company went through the SOSV HAX hardware accelerator and Morgan Stanley’s Inclusive Ventures Lab. To keep up the pace, it is now raising a $5 million late-seed round, a funding stage that comes just before a larger Series A.

Danu’s next aim is a sturdier, smaller and more flexible HERO. Once it is portable enough, Ma wants it to work outside dedicated waste plants as a stand-alone recycling station.

She names events, shopping centers, hospitals and airports as likely places, where waste could be sorted right where it is thrown away. Sorting at the source would mean less mixed waste trucked to plants and cleaner material from the start, which she sees as the real way to change how packaging waste is handled.

There is no date or price for a smaller model yet. The company’s longer-term goal is for multi-arm AI sorting systems to become standard in major recycling facilities around the world within about five years.

For most people, the effect is indirect but real. If robots like HERO raise the share of material that gets recovered, more of what you rinse and put in your bin has a chance of becoming something new instead of landfill.

Better sorting can also help local councils, which often pay for collection and processing. When plants recover more material and get fined less for contamination, recycling programs are cheaper to run.

Until those robots are common, your own habits still matter most. Keep food and liquids out of recyclables and follow your local council’s sorting rules, because a contaminated load is far more likely to end up in the trash, robot or no robot.