TeknTrash
Last reviewed
May 9, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v5 · 2,917 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
May 9, 2026
Sources
No citations yet
Review status
Needs citations
Revision
v5 · 2,917 words
Add missing citations, update stale details, or suggest a clearer explanation.
| TeknTrash | |
|---|---|
| General information | |
| Full name | TeknTrash Robotics (legal entity: TeknTrash Inc Ltd) |
| Former name | Synthetic Data Inc Ltd (2018 to 2022) |
| Type | Private limited company |
| Founded | 16 November 2018 |
| Founder | Al Costa (CEO) |
| Headquarters | 37th Floor, 1 Canada Square, Canary Wharf, London E14 5AA, United Kingdom |
| Industry | Robotics, waste management, recycling |
| Products | ALPHA, HoloLab, Stipra (also branded as ProduTrak) |
| Subsidiary | Titan Rui Technology (China, Suzhou) |
| Investors | Google for Startups |
| Website | tekntrash.com |
TeknTrash (legally registered as TeknTrash Inc Ltd and trading as TeknTrash Robotics) is a United Kingdom-based robotics research company that develops artificial-intelligence-driven solutions for waste handling, recycling, and post-sales consumer analytics. The company is headquartered at 1 Canada Square in London's Canary Wharf district and was founded by Al Costa, who serves as chief executive officer. Its product portfolio combines a humanoid robot named ALPHA, a cloud-based monitoring platform called HoloLab, and a consumer-data system originally branded Stipra and later rebranded as ProduTrak.[1][2][3]
TeknTrash positions itself around the principle that waste handling is dangerous, unsanitary, and degrading work that should be performed by machines rather than people. Its commercial activity has centred on a public pilot launched in April 2025 with the British waste-management firm Sharp Group at a recycling plant in Rainham, East London, where ALPHA is being trained to sort mixed recyclables on conveyor belts. In November 2025 the company opened a Chinese base in Suzhou Industrial Park to support assembly, testing, and Asian commercial expansion.[4][5][6]
The legal entity behind TeknTrash was incorporated in England and Wales on 16 November 2018 under the name Synthetic Data Inc Ltd, with company number 11681259. The company was renamed TeknTrash Inc Ltd on 23 May 2022. According to UK Companies House records, the firm is registered as a private limited company at 1 Canada Square in London and remains active, with business activities listed under information-technology consultancy and engineering consulting.[7]
In external profiles maintained by data providers such as Crunchbase and CB Insights, TeknTrash's founding year is commonly cited as 2019, the year in which the trading operation began under the TeknTrash brand. The company's earliest commercial product was Stipra, an AI-and-analytics platform that tracked the disposal of consumer-packaged goods at recycling points and rewarded people for proper recycling. TeknTrash describes Stipra as the first service of its kind to combine post-consumption product data for brands with reward-based recycling for consumers.[1][3][8]
Al Costa is a serial entrepreneur whose career spans the United States, the United Kingdom, Spain, and Brazil. He holds a Bachelor of Arts in biology from Saint Leo University and a master's degree in computer engineering from the University of South Florida. He has stated that this dual background informs his approach to robotics, which combines biological and engineering perspectives.[2][9]
Before founding TeknTrash, Costa established several other ventures and sold his first company to a NASDAQ-listed firm. He is also chief executive of Alkol Biotech, a Spanish company that has been described as the only commercial supplier of sugarcane in Europe and that produces a hybrid sugarcane crop called EUnergyCane in Motril, in southern Spain. Costa is the author of five books covering information technology, biofuels, and a science-fiction novel, and has taught Big Data and machine learning at EAE University in Spain. He has also acted as a project advisor on biotechnology, AI, and high technology for European Union agencies.[2][9][10]
TeknTrash's pivot from analytics-only products toward humanoid hardware became public in April 2025 with the announcement of the Sharp Group pilot. The company adopted the position that conventional robotic arms used in recycling plants are limited because they are stationary and rely on suction grippers, and it began designing a mobile humanoid platform that could perform a broader range of waste-handling tasks. The transition built on data and computer-vision know-how developed during the Stipra and ProduTrak programmes.[4][11]
ALPHA, short for Automated Litter Processing Humanoid Assistant, is TeknTrash's flagship product and a humanoid robot designed for waste-sorting tasks on recycling lines. The hardware is built by the Chinese manufacturer RealMan Robotics and is then adapted by TeknTrash for recycling environments through software, vision, and training systems. Although TeknTrash has not published full mechanical specifications, RealMan's RM75 family of seven-degree-of-freedom arms used in humanoid platforms have a 5 kilogram payload, a 610 millimetre working radius, and a self-weight of about 7.8 kilograms.[5][12][13]
ALPHA is presented as an alternative to stationary sorting arms. According to TeknTrash, the robot is designed for mobility, dexterity, and perception, and is intended over time to take on additional duties such as carrying bins to collection trucks and lifting weights inside processing plants. Multiple ALPHA units can be coordinated to operate together along rails inside a sorting facility.[4][5][14]
Key design choices reported by TeknTrash and its press partners include:
| Subsystem | Approach |
|---|---|
| Vision | Hyperspectral cameras placed at the start of the conveyor belt, intended to track items earlier and more accurately than the standard RGBD cameras used by competing solutions. |
| Manipulation | Dexterous grippers trained to mirror human hand and finger motion, presented as a higher-capture alternative to suction-based pickers. |
| Mobility | Autonomous travel along guide rails with multi-unit coordination across a sorting line. |
| Compute | A cloud-first architecture that offloads heavy perception and planning work to remote servers, leaving lightweight, energy-efficient hardware on the robot itself. |
| Training | Worker demonstrations captured in Meta Quest 3 virtual-reality headsets and processed through imitation learning. |
| Stack | NVIDIA Isaac Lab for training and NVIDIA GR00T for inference, with worker data stored in LeRobot format. |
HoloLab is TeknTrash's cloud-based monitoring platform and the data backbone for ALPHA's training. According to TeknTrash, the system tracks disposed products in real time at waste facilities, exposes the resulting information through interactive maps and dashboards, and exports the records for downstream analytics. The platform pulls video from cameras that use the RTSP streaming format and supports input from mobile devices and head-mounted displays such as the Meta Quest 3.[14][15]
HoloLab is offered to three groups of customers: recycling plants that want to lift recovery rates and identify operational inefficiencies, fast-moving consumer-goods (FMCG) companies that want post-sales information about how their packaging is disposed of, and robotics teams that want training data. For the robotics use case, HoloLab records operator movements, including posture, hand and finger articulation, and synchronised video, and saves the result in the LeRobot format used by Vision-Language-Action models such as NVIDIA's Isaac GR00T.[14][15]
Stipra, also branded as ProduTrak, predates ALPHA and HoloLab. It is described by TeknTrash as the first system to provide companies with real-time consumption data captured at the moment of disposal. Inputs come from a consumer mobile application, cameras placed at recycling plants, and reverse-vending machines. Brands receive analytics about how their products are actually being thrown away, and consumers can earn rewards for correct recycling behaviour.[1][3]
The TeknTrash press materials note that the underlying recognition stack uses a four-stage pipeline based on semantic segmentation, the YOLO object-detection family, generative adversarial networks, and reinforcement learning, and that the company has received support that includes US$120,000 of cloud storage and processing from IBM in addition to its participation in the Google for Startups campus in London.[3][8]
In April 2025 TeknTrash announced a public pilot with Sharp Group, an environmental-services company also known as Sharp Skips. Sharp Group is a third-generation family business based at Unit 6 Albright Industrial Estate, Ferry Lane, Rainham (postcode RM13 9BU), in East London. The firm was founded in 1983, operates a fleet of 38 ULEZ-compliant vehicles, runs an on-site solar installation, and reports diesel savings of around 1,900 gallons per month from its fleet. Its finance director, Chelsea Sharp, is the granddaughter of founder Tom Sharp.[6][16]
The Rainham facility processes around 2,800 tonnes of waste per week, equivalent to roughly 280,000 tonnes of mixed recycling per year. It currently runs from 7 a.m. to 7 p.m. and has capacity for 24-hour operation, but Sharp Group has stated that staffing three shifts has not been possible because the labour market is unwilling to take on the work conditions involved. The line uses around 24 agency workers at a time and has reported staff turnover of about 40 per cent per year, which TeknTrash and reporting outlets have used as a primary justification for automation.[6][11][17]
The pilot is structured as a six-month period of data capture in which Sharp Group operatives wear Meta Quest 3 headsets while sorting waste. TeknTrash's HoloLab software records the workers' finger positions and the video they see, then sends both feeds to Google Cloud servers. The captured demonstrations are converted into LeRobot-format datasets and used to train ALPHA through Isaac Lab, with the resulting models executed at inference time using NVIDIA's GR00T framework. Once the pilot concludes, TeknTrash plans to use the Rainham facility as a launchpad for a wider rollout across the United Kingdom and an additional 1,000 plants across Europe over the following 24 months.[4][5][14][16]
Sharp Group has framed the partnership as augmenting rather than displacing its workforce. According to interviews with Chelsea Sharp, existing staff are intended to move into maintenance and supervisory roles as ALPHA is deployed.[11][17]
Before the humanoid pilot, TeknTrash deployed a computer-vision system at a tire recycling plant in Madrid run by SACYR S.A., a Spanish infrastructure operator. According to TeknTrash, the facility is the largest tire recycling plant in Europe and processes around 20,000 tonnes of used tires per year, broken down into approximately 3,000 truck tires and 17,000 car tires. About 40 per cent of the rubber is granulated for products such as playground flooring, while the remaining 60 per cent is used as fuel for cement plants.[18]
The plant uses a two-stage cutting process in which steel teeth from the first machine occasionally pass through to the second, damaging equipment and stopping production. TeknTrash installed four continuous-recording cameras above the first cutter and applied a YOLO-based detection model that captures images every second, identifies metal contamination, and triggers an immediate stop signal through the plant's IoT control system. The case study is presented by TeknTrash as an example of how its perception stack can be deployed in industrial settings without humanoid hardware.[18]
TeknTrash's approach to robotics emphasises specialisation rather than general-purpose humanoids. In an interview with Robot Magazine, Costa argued that robots should follow the same logic as human workers, building expertise in a single domain such as waste handling rather than aiming to perform many unrelated tasks. The company's design choices for ALPHA reflect that orientation: the hardware is configured for sorting and material handling, the perception system targets recyclables on conveyor belts, and the training pipeline is anchored in real demonstrations from the recycling industry rather than generic motion data.[2]
A second design pillar is cloud-first computation. TeknTrash participates in the One6G Association, a consortium that researches future cellular and edge connectivity, and has been described by Google Cloud as an example of cloud robotics. Heavy perception and planning workloads run on remote servers, while the robot itself is intended to be relatively lightweight and energy-efficient. According to TeknTrash, this architecture is what allows multiple ALPHA units to be coordinated across a sorting line and to be updated centrally as new training data arrives.[2][8]
A third pillar is data collection at scale through HoloLab. TeknTrash has stated publicly that an effective humanoid waste-sorting system requires extensive demonstrations and that training will take months, in part because the diversity of materials and contamination conditions in real plants is much greater than in laboratory settings. The Rainham pilot is presented as a means to build that dataset, with the eventual goal of replicating the same software stack across 1,000 European plants connected to the cloud.[5][11][17]
TeknTrash's commercial pitch is grounded in publicly cited statistics about labour conditions in the waste-management sector. In its press release covering the Sharp Group pilot, the company referenced 2018 to 2019 figures showing that the recycling and waste-management industry reported a worker illness rate of 4.5 per cent compared with 3.1 per cent across all industries, and a non-fatal injury rate of 3.4 per cent compared with 1.8 per cent. Fatal-injury rates in the same period were reported at around 17 times the cross-industry average.[4]
Reporting in The Next Web, BM Magazine, and other outlets covering the pilot has placed TeknTrash within a broader picture of British recycling firms struggling with labour shortages. According to those reports, the sector experiences staff turnover of around 40 per cent per year and a fatality rate roughly eight times the national average, with work-related injury and illness rates around 45 per cent higher than other sectors and average worker tenure of about 30 months. Human pickers on conveyor lines typically achieve 30 to 40 picks per minute, with single-stream recycling streams reporting contamination rates of around 25 per cent.[6][11][17]
TeknTrash uses these figures to argue that automation is necessary not only to lift purity rates and profitability for plant operators but also to remove human workers from a hazardous task. The company's stated long-term goal is to redeploy human staff into roles such as supervising robots, maintaining equipment, and managing data pipelines, while reserving the actual sorting work for ALPHA units.[4][6]
The United Kingdom is TeknTrash's home market and the location of its registered office in Canary Wharf. The Rainham pilot is positioned as the launch site for the British rollout, and TeknTrash has cited the country's labour shortages and waste-management costs as commercial drivers.[6][7]
In November 2025 TeknTrash opened a base inside Suzhou Industrial Park, where it operates under the local business name Titan Rui Technology. The Suzhou facility is intended for robot assembly, testing, and local research and development, and it acts as the company's regional headquarters for Asian markets. TeknTrash has described the move as an effort to combine European engineering with Chinese manufacturing, and has cited support from the Suzhou Industrial Park administrative committee covering technical validation, local recruitment, and supply-chain optimisation.[19]
The Suzhou base is also expected to support development of the next generation of ALPHA, which is being designed to handle waste in both domestic and industrial settings. TeknTrash has indicated that future demonstrations of ALPHA in China will be carried out with local partners, and Costa has publicly considered relocating personally to China to pursue the company's Asian strategy.[2][19]
TeknTrash's earlier deployments in Spain include the Madrid tire recycling project at SACYR's plant. The company has also referenced operations and partnerships across the United States, Brazil, and other European markets through Costa's broader entrepreneurial network, including Alkol Biotech in Spain. Press coverage has identified Sharp Group's Rainham facility as the European reference site for the planned 1,000-plant rollout.[2][9][18]
TeknTrash is listed by external data providers as an early-stage company at the incubator and accelerator level. Crunchbase and CB Insights both record Google for Startups as an investor, and the company has stated that it is part of the Google for Startups campus in London. TeknTrash has additionally reported receiving US$120,000 in cloud storage and processing credit from IBM as well as recognition from Google Cloud as an example of cloud robotics. Press materials list coverage by Euronews, Startup Magazine, Ecommerce News, and the World Economic Forum.[3][8]
The company describes its underlying recognition system as covered by patentable technology, organised in a four-stage pipeline that uses semantic segmentation, YOLO, generative adversarial networks, and reinforcement learning. TeknTrash's published material does not currently disclose total funds raised, headcount, or formal investor rounds beyond its participation in the Google for Startups programme.[3][8]