Edinburgh-based Danu Robotics has raised $5 million in late-seed funding to bring its AI-powered recycling robot, H.E.R.O., to market, according to TechCrunch AI. Founder Amy Ma has spent six years on the idea. Danu says it already has $500,000 in signed contracts and more than 200 potential customers in its sales pipeline. The company is also heading to TechCrunch Disrupt’s Startup Battlefield 200 in San Francisco on October 13-15.
The Quick Take
- What: A recycling-sorting robot that picks items with a pincer claw rather than suction, and runs on software that AI keeps improving.
- Who: Danu Robotics, founded by former London bank software developer Amy Ma.
- Money: $5M late-seed round, $500K in signed contracts, letters of interest from two large customers.
- Market: Danu puts the recycling-sorting market across Europe and North America at about $20 billion.
- Pitch: Ma estimates $485,000 in added revenue per site, against a $160,000 upfront cost and $24,000 a year in maintenance.
How a Bank Recycling Bin Started a Robotics Company
The origin story is simple. Ma worked in a well-funded London office and noticed that the building’s carefully labeled recycling bins were being emptied into the trash. “Even though they have tons of money, there’s all these different recycling bins that just end up in the same place,” she told TechCrunch.
When she dug in, she found the real bottleneck. At most facilities, recycling still depends on people sorting waste by hand off moving conveyor belts. The work is slow and expensive, and that cost decides what gets recovered and what goes to landfill. “I saw the technology and thought, we can do so much better,” Ma said.
What Makes H.E.R.O. Different
Robotic waste pickers already exist. Competitors like Glacier and RecycleEye sell suction-based arms, and those have become the standard approach. Danu stands out in two ways:
- A pincer claw instead of suction. Suction works best on flat, clean surfaces. A gripper can, in principle, deal better with the crushed, dirty, oddly shaped items that actually come down a sorting line.
- Software that keeps getting better. Danu uses AI to keep improving the robot’s software. In practice, that means the vision and picking models should get better at spotting materials as they see more real waste.
The second point matters more. In AI-driven robotics, the hardware is often easier to copy than the data flywheel behind it. Every robot on a sorting line collects examples, and those examples feed better models. If Danu can turn its early deployments into steady model improvements, that advantage compounds over time.
Why the Economics Could Work
Recycling robots have a clearer sales pitch than most AI hardware. Recycling centers earn money from the material they recover, so a robot that pulls more sellable plastic, metal or cardboard off the line pays for itself directly.
Ma’s numbers make that case: about $485,000 in extra revenue per site against $160,000 upfront and $24,000 a year to maintain. TechCrunch didn’t say what period the revenue estimate covers, and these are company projections, not audited results. Still, the math points to a short payback, and Ma calls it significantly cheaper than rival systems.
What stands out is that Danu sells on ROI rather than sustainability messaging. Facility operators work on thin margins, and a clear return is much easier to approve than a green-branding exercise.
What Comes Next
H.E.R.O. units are already shipping, so Ma’s next goal is a sturdier, smaller and more versatile robot. A portable version would take recycling beyond dedicated facilities.
“Once it’s smaller, it can be used as a stand-alone recycling solution for events, for shopping centers, hospitals, or airports to recycle their waste at the source,” Ma told TechCrunch.
That’s the bigger bet here. Sorting waste where it’s produced, before it’s mixed and contaminated, could lift recovery rates far more than better sorting at the end of the line.
What to Watch
- Pipeline conversion: 200+ prospects is a strong signal, but $500K in signed contracts is still early. Watch how many of them become paying deployments.
- Performance data: Claims that the claw beats suction need public picking-rate and accuracy numbers to back them up.
- Portability timeline: A compact unit would open new markets, but shrinking industrial robotics is hard and expensive.
Danu is a good example of a wider trend: physical AI moving into unglamorous industries where the ROI is easy to measure. Expect more attention on the company after its Disrupt appearance next week. TechCrunch AI has the full story.