← All articles

Mikhail Polyakov

Danu Robotics has spent six years building a waste-sorting robot: what its founder promises and how to check such a payback estimate

Level: BeginnerReading time: 5 min

Topics: Robots, Industry

Above a conveyor carrying bottles, boxes and cans, a mechanical pincer claw with amber lighting lifts a bottle and places it into one of several glowing single-color bins

Short answer: according to a TechCrunch article of October 9, 2026, the Edinburgh startup Danu Robotics has spent six years making H.E.R.O., a robot that sorts recyclable waste. It picks up items with a pincer claw rather than a suction system, and the company uses AI to keep improving the robot's software. The startup has raised $5 million, has signed contracts worth $500,000 and letters of intent from two large customers, and has more than 200 customers in its sales pipeline. Founder Amy Ma estimates that a working site will earn $485,000 in additional revenue against an initial investment of $160,000 and maintenance costs of $24,000 a year. This is the company's own estimate, and the article does not state the period over which this revenue accrues. For a business owner, the story is useful as an example of how a supplier calculates payback and what to check in such a calculation. Below are the facts and five questions.

What is known

All the facts below come from the TechCrunch article; the economic figures are the founder's estimates.

  • Where the idea came from. Amy Ma worked for years as a software developer at a London bank and noticed that the recycling bins in the building were being emptied into the general trash. She says that the economics of recycling often come down to people sorting streams of waste by hand.
  • What the robot is. H.E.R.O. is designed, by the company's intent, to sort recyclable waste more cheaply and quickly. It picks up items with a pincer claw, not the suction that many competitors use. The more important difference, in the outlet's view, is the business approach: Danu uses AI to keep improving the robot's software.
  • What already exists. Signed contracts worth $500,000, letters of intent from two large customers, more than 200 customers in the sales pipeline and $5 million raised at the late seed stage. The article puts the market for recycling sorting in Europe and North America at $20 billion.
  • Competitors. There is already a sizable market for waste-sorting robots, especially suction-based ones; the companies Glacier and RecycleEye are named.
  • The payback estimate. Ma says that recycling centers make money from the materials they sort out, so a working system has a quick return on investment. Her calculation: $485,000 in additional revenue against $160,000 of initial investment and $24,000 a year in maintenance.
  • Plans. To make the robot sturdier, smaller and more versatile, so that it can be used outside dedicated facilities: at events, in shopping centers, hospitals and airports. These are plans, not results.
  • Upcoming event. Danu is taking part in the Startup Battlefield 200 competition at TechCrunch Disrupt, October 13 to 15, in San Francisco.

What the article lacks: independent measurements of the robot's work, the period over which the $485,000 is earned, and data on sorting accuracy. I do not name them.

Five questions for a robot supplier

These are my conclusions from what I read, not the words of Danu or TechCrunch. They are written for buying a robot that is supposed to bring additional revenue. For equipment that pays back through lower costs, these questions are not considered.

  1. Over what period is the stated additional revenue earned? In this story the period for $485,000 is not given, and without it the figure cannot be compared with the investment.
    • Who does it: whoever is purchasing.
    • How to check: in the commercial offer, every sum of revenue or benefit has a period (month, year) and assumptions: stream volume, share sorted, price of materials.
  2. What assumptions is this calculated on? The benefit depends on your volumes, the composition of your raw material or workpieces, and prices.
    • Who does it: whoever is purchasing, together with a process engineer.
    • How to check: compare the supplier's assumptions (volume, composition, prices) with your real data for the last three months. If your data differ, ask for a recalculation.
  3. Who has confirmed the result besides the supplier? Here the article gives the founder's estimate, not an independent measurement.
    • Who does it: whoever is purchasing.
    • How to check: the supplier has a customer willing to talk about results, or a trial period on your site with measurement before and after.
  4. What does the maintenance sum include? In the source it is $24,000 a year.
    • Who does it: whoever is purchasing.
    • How to check: the contract lists what is included (software updates, spare parts, visits) and what is paid separately.
  5. What happens if the robot does not deliver what was promised? A $500,000 contract and letters of intent do not guarantee results.
    • Who does it: a lawyer or whoever is purchasing.
    • How to check: the contract has conditions for refund or termination tied to a measurable indicator.

A calculation with illustrative numbers

Suppose a supplier names $485,000 of additional revenue and does not state the period. This is revenue, not profit: you have other costs too, for example staff, electricity and downtime. For comparison, let us take only the sums named in the source: investment of $160,000 and maintenance of $24,000 a year. If the $485,000 is per year, annual revenue exceeds the investment and maintenance together ($184,000) by $301,000, before other costs. If, however, the $485,000 accrues over five years and, say, evenly across the years, revenue will be $97,000 a year, and in the first year it will be $87,000 less than the $184,000 of named costs, again before other costs. The difference between the two cases shows why the period must be found out before signing, and why the supplier's assumptions should be checked against your own stream. The numbers are illustrative: take your own from the offer.

When this does not concern you

If you are not in waste sorting, you do not need the robot itself. The questions about the period of the benefit and independent confirmation can be asked at any purchase where the supplier promises payback; the rest will have to be rewritten for your equipment.

Summary

According to TechCrunch, Danu Robotics has spent six years making the H.E.R.O. robot for sorting recyclable waste, has raised $5 million and names an estimate: $485,000 in additional revenue against an investment of $160,000 and $24,000 a year in maintenance. The period for earning this revenue and independent measurements are not named in the article. Such an estimate is worth checking by period, assumptions, independent confirmation, what maintenance includes and the conditions if the robot underdelivers.

I work on AI agents and automation. If you want to look at my projects or discuss your own task, visit my portfolio.

Sources