The most honest piece of climate tech is the public trash can. It's where good intentions go to die, a jumble of half-empty coffee cups and crumpled foil that guarantees a recycling stream will be contaminated and landfilled. For nearly a decade, CleanRobotics has been trying to fix that moment of failure with a robot. Its flagship product, TrashBot, is a smart bin that closes a door after you toss something in, images and weighs the item, and sorts it internally into landfill or recycling. It's a simple premise with a stubbornly difficult execution, aimed at the chaotic waste streams of airports, stadiums, and hospitals [Perplexity Sonar Pro Brief, retrieved 2024].
The airport as a proving ground
CleanRobotics has found an ideal early customer in the modern airport. It's a controlled, high-footfall environment with a public sustainability mandate and a waste management headache. The company's most cited case study comes from an unnamed airport, where a TrashBot unit sorted nearly 2,500 items with 96% accuracy and abated an estimated 307 pounds of carbon dioxide equivalent [cleanrobotics.com, Unknown]. The real value for facility managers, however, is likely in the data. Each unit runs on Sierra Wireless cellular modules, feeding a monthly software service that tracks material types and volumes [Semtech / Sierra Wireless, 2020s]. For an operations team, knowing the composition of waste is the first step toward reducing its cost and its carbon footprint.
The company lists a roster of pilot and deployment partners that reads like a who's who of large, complex facilities. Dallas Fort Worth International Airport, The Port Authority of New York and New Jersey, Children’s Hospital Los Angeles, and Google are among the names cited [aroptions.com, Unknown] [sbir.gov, Unknown]. This suggests a sales motion that targets institutional sustainability and facilities departments, a longer cycle but with the potential for multi-unit deployments.
Hardware, software, and a shifting team
Building a robot that can reliably sort wet, misshapen trash in public spaces is a full-stack hardware and AI challenge. CleanRobotics has developed three models,TrashBot, TrashBot Slim, and TrashBot Zero,to accommodate different volumes and waste streams [technical.ly, retrieved 2024]. The technical lineage is traceable to Pittsburgh's robotics ecosystem. Co-founder and former CTO Koushil Sreenath was an assistant professor in mechanical engineering at Carnegie Mellon University and is now a professor at UC Berkeley [technical.ly, retrieved 2024] [hybrid-robotics.berkeley.edu, Unknown].
The leadership structure, however, appears to have evolved. Sources name multiple co-founders, including Charles Yhap (now CEO), Thomas Askey, Tanner Cook, and Vaish Krishnamurthy (a former CEO) [Tracxn, retrieved 2026] [technical.ly, retrieved 2024]. The company currently reports having between 11 and 50 employees [LinkedIn, retrieved 2024]. This kind of team evolution is not uncommon for a hardware startup founded in 2015, but it underscores the long road from prototype to product.
Seed (Unknown) | 0.1 | M USD
SBIR Grant (Unknown) | 0.4 | M USD
Series A (Aug 2022) | 4.5 | M USD
A crowded field of sorting robots
The problem CleanRobotics is solving is obvious, which is why it's not alone. The competitive set is global and well-funded.
| Company | Primary Focus | Notable Differentiator |
|---|---|---|
| AMP Robotics | Post-collection sorting at MRFs | High-speed, AI-guided robotic arms on conveyor belts. |
| Recycleye | Waste analytics & robotic picking | Computer vision for audit and quality control. |
| Glacier | Robotic sorting at MRFs | Lower-cost, AI-powered robotic sorter. |
| CleanRobotics | Point-of-disposal sorting | Integrated bin for public spaces, focuses on prevention. |
CleanRobotics's wedge is distinct: it intervenes at the point of disposal, aiming to prevent contamination before it happens. Competitors like AMP Robotics typically operate in material recovery facilities (MRFs), sorting waste after it's already been collected and mixed. This is a key strategic fork. Is it better to build a perfect sorter for the back end, or a prevention tool for the front? CleanRobotics is betting on the latter, arguing that clean streams are cheaper to process. The risk is that its hardware must be durable enough for public use and economical enough to deploy at scale, a tougher unit economics problem than a stationary robot in a controlled factory.
The unit economics of a single bin
Let's run a back-of-the-envelope check. Assume a TrashBot unit costs a facility $15,000 (a speculative figure for hardware of this complexity). If it replaces the labor needed to hand-sort bins and avoids contamination fees from recycling processors, the payback period could be reasonable for a high-volume location. The 307 pounds of CO2 abated in the airport case study is a start, but the climate math only works if the bin consistently diverts recyclables for years. The real test is whether the data service,tracking what's thrown away,creates enough operational savings to turn the robot from a sustainability showcase into a must-have operations tool.
For CleanRobotics to graduate from pilots to widespread adoption, it must beat the incumbent: the humble, dumb bin paired with cheap manual labor. Its argument is that labor is getting more expensive and sustainability mandates are getting stricter, making the automated alternative inevitable. The next twelve months will likely show if that logic holds in purchase orders. The company will need to move beyond undisclosed pilots to announced multi-unit deals with the airports and hospitals on its list. Another funding round may also be on the horizon to scale production; its last disclosed raise was a $4.5 million Series A in August 2022 [Tracxn, retrieved 2026].
Sources
- [Perplexity Sonar Pro Brief, retrieved 2024] Brief on CleanRobotics product and target market
- [cleanrobotics.com, Unknown] Case study on airport deployment and accuracy metrics
- [Semtech / Sierra Wireless, 2020s] Customer story on TrashBot connectivity
- [aroptions.com, Unknown] List of customer deployments
- [sbir.gov, Unknown] List of pilot customers including Google and airports
- [technical.ly, retrieved 2024] Article on company history and product models
- [hybrid-robotics.berkeley.edu, Unknown] Academic profile of Koushil Sreenath
- [Tracxn, retrieved 2026] Funding round data and founder list
- [LinkedIn, retrieved 2024] Company headcount estimate