UK recycling company uses Chinese-made robot as waste sorting sector faces 40% annual job losses and 8 times higher death rate

The recycling industry has a labor problem that no amount of hiring can solve. Employee turnover in waste sorting facilities is 40 percent per year. The death rate is eight times the national average for all industries. Work-related injuries and illnesses are 45 percent higher than other sectors. The job involves standing next to a speeding conveyor belt, dragging shoes, concrete blocks, VHS tapes, and occasionally guns in a stream of mixed debris, in conditions so dusty and noisy that the people who do it rarely stay long enough to succeed. The industry has experimented with higher wages, flexible shifts, and agency workers. Nothing has changed in the basic calculation: the work is dangerous, unpleasant, and physically demanding, and the people who do it leave as soon as they find something else. In a skip yard in east London, a family-run waste disposal company has concluded that the answer is not a better recruitment strategy. It is a trained humanoid robot designed to replace it.

Robot

The Sharp Group processes 280,000 tonnes of mixed recycling a year at its site in Rainham, east London, using 24 agency workers on high-speed conveyor belts. The company, founded by Tom Sharp and now run by the third generation of the family, has released a humanoid robot called Alpha, built by RealMan Robotics in China and adapted for recycling by the British company TeknTrash Robotics.

Alpha stands in line as a human worker. That’s the point. TeknTrash founder Al Costa says the humanoid form factor allows the robot to fit into existing plant structures without requiring the facility to be redesigned around it. Alternatively, companies like Colorado-based AMP and California-based Glacier have pursued, purpose-built planning systems using robotic arms, air jets, and AI vision. Those programs work, but they require new resources or expensive reimbursements. A humanoid that can pick up where a human stopped and do what a human did is, in theory, a cheap and quick way to automate hundreds of small recycling plants that can’t be rebuilt.

Alpha is not working yet. When the BBC visited, he was in a training session, guided by arm movements while a factory worker at his side wearing a Meta Quest 3 VR headset, recorded his editing movements to show what a successful pick-up looked like. TeknTrash’s HoloLab system feeds data from multiple cameras to train a robot in two parallel tasks: identifying what’s on the belt and physically picking it up. Thousands of objects pass through the system every day, creating millions of data points. Costa is candid about the timeline. “The market thinks that these robots are ready to be worn, all you need to do is connect them to the mains and they will work well. But they need extensive data to be effectively useful.” The training will take months. TeknTrash plans to use the same system at 1,000 plants in Europe, all connected to the cloud, but that ambition depends on Alpha learning to filter reliably at one plant first.

The competition

The humanoid approach is unusual. The auto recycling market is dominated by companies that have taken a different approach. Sereact raised $110 million in April to standardize AI that enables any industrial robot to adapt to conditions across materials and manufacturing, reflecting a broader investment perspective that the value is in the software layer, not the physical form. AMP, a Colorado-based filtration company, has raised $91 million in its Series D and now operates three of its plants while providing AI-enabled filtration equipment to more than 100 facilities worldwide. Its system uses jets of air to direct material into pieces at eight to ten times the speed of a human worker. CEO Tim Stuart, a former CEO at Republic Services, describes this approach as very different from trying to replicate human movement: building filtering intelligence into the system and designing the physical infrastructure around it.

Glacier, an Amazon-backed California startup founded by Rebecca Hu-Thrams, has taken a middle ground: embedded robotic arms controlled by AI vision systems that can be installed in existing facilities without a full rebuild. The company has raised $16 million by 2025, considered recycling for 1 in 10 Americans, and was named to TIME’s Best Inventions list. Hu-Thrams emphasizes that the Glacier system is designed to work for rural areas on tight budgets, not just large urban plants. AI learns from over a billion programmed objects, continuously improving. Waste diversity is a key technical challenge. “Sometimes the beer will be spraying liquid everywhere, threatening the equipment,” Hu-Thrams said. His clients encountered hand grenades and guns in the sorting line.

Industrial logic

Siemens used Nvidia’s humanoid robot in a live factory environment in January, picking totes from storage stacks and moving them on conveyor belts in a two-week trial. Experiments have shown that humanoid robots can work in real industrial settings, but also reveal a gap between controlled demonstrations and continuous production use. The recycling situation is complicated. Factory floors are orderly and predictable. A recycling conveyor belt carries a random assortment of materials at varying speeds, many of which are wet, broken, or clumped together. A humanoid robot capable of reliably sorting waste can, by definition, perform many of the sorting and sorting tasks of a factory. The recycling line, in engineering terms, is one of the most difficult areas to automate.

Tesla is eyeing mass production of its Optimus humanoid robot from the Shanghai Gigafactory, with more than 1,000 Gen 3 units already deployed across Tesla’s facilities and full-scale production scheduled for 2026 to 2028. rotate, and place objects of different shapes and weights. That ability is exactly what recycling requires. Alpha’s manufacturer, RealMan Robotics, is part of China’s robotics ecosystem that produces humanoids at price points that Western manufacturers can’t match. The geopolitics of humanoid robots mirrors the geopolitics of semiconductors: the hardware is increasingly Chinese, the software layer is contested, and the areas of use are global.

Economics

The financial case for automating recycling is straightforward. A sorting line job costs around £25,000 to £30,000 a year in the UK including agency fees, and leaves on average 30 months at current rates of pay. The costs of constantly hiring, training, and relocating workers add up to a structural drag on the edges of an industry where margins are already thin. A robot that works 24 hours a day, seven days a week, without holidays, sick days, and the risk of injury, changes the unit economy of every ton processed. “The appeal of the humanoid is that you can put it here and it stays here,” said Chelsea Sharp, the plant’s finance director and granddaughter of the founder. “It will choose all day, 24 hours a day, seven days a week.”

Accenture has invested in General Robotics to integrate factory robots with integrated AI, part of a broader pattern in which technology consulting firms are building software infrastructure to manage dozens of industrial robots across multiple sites. The recycling industry is the first discovery of nature because its labor economy is very poor in production: very high turnover, very high injury rates, and less desirable working conditions. If default works here, it works almost anywhere. Professor Marian Chertow of Yale University describes this change as inevitable and necessary: ​​robots and vision systems driven by AI offer the greatest potential to improve material recovery, worker safety, and economic competitiveness in recycling.

Employees

A question that automation always raises, and one that the recycling industry cannot avoid, is what happens to people whose jobs are taken over by robots. Sharp Group employs 24 agency workers on its sorting lines. If Alpha and its successors are able to match the levels of screening people, AMP systems already exceed by a factor of eight to 10, those 24 positions are maintenance and supervisory roles. Chelsea Sharp says the plan is to improve the staff available to maintain and guide the robots, removing them from the dust, noise, and physical hazards of the conveyor belt. The narrative is common across all automating industries: dangerous jobs are being eliminated, workers are being retrained, and new roles are becoming better. Whether that happens in practice depends on whether the company invests in retraining and whether employees have the skills and desire to transition into technical maintenance roles. In an industry with a 40 percent annual turnover, most of the current workforce will be gone before the robots are fully operational.

What’s happening in Rainham is a small version of the revolution that’s happening in every industry where the work is too dangerous, too unpleasant, or too poorly paid to keep workers. The recycling industry consumes the waste of the entire economy, and has done so for decades using the cheapest labor available in the worst conditions available. The humanoid robot on Sharp Group’s sorting line has yet to replace its human sidekick. But next to the person will not stay. The industry standard of 40 percent is not a hiring failure. It’s a sign that work was never meant for people in the first place, and an acknowledgment that technology has finally arrived.