AI in waste/recycling management – smarter city services

Waste and recycling get’s smart – moving towards smarter cities!

Rubicon Overview

              AI and machine learning can help make difficult, time-consuming, and labor-intensive processes easier to do and potentially, economically viable. Rubicon Global uses AI to improve upon digital waste and recycling solutions for their customers. Rubicon operates globally and offers several different product solutions for their customers. Some include smart bins that notify the collection company when the bins are full and need to be picked up and data tracking systems to understand the true recycling rate in municipalities. As well, the company has entered into a partnership with Samsara to help with route optimization and other service optimization which are described in more detail below.

Rubicon sees 3 main technology trends affecting the waste and recycling sector today. The first, is the increasing use of IoT for asset tracking. The second, is route optimization. And the third is better recycling sorting using technology.

Creating Value

Rubicon helps cities and waste management collection companies create value by using AI, machine learning, and a vast array of sensors. To create this value, Rubicon has sensors mounted on various parts of the trash truck (including the breaks, steering wheel, and mechanical trash compacting parts). There are also cameras mounted on the truck that take in video feed. The tape is analyzed by AI-enabled image-recognition software to identify things on the road. Rubicon offers haulers an app that helps communicate the insights to drivers in real time and also a dashboard offering to help company management plan based on the insights.

They take this data and use AI-enabled analytics to reduce costs for cities by strategically picking up garbage only when needed which reduces fuel consumption, increasing collection rates which reduces redundant trash runs to pick up “missed” garbage collection, and also provides information directly to the city about the real-time status of the road conditions, graffiti, tree branches, and trash patterns among other data points. This service also helps municipalities better plan for reoccurring events such as parades, concerts, and protests to estimate the types of clean up they will need in advance instead of having to react to the clean-up requirements after the fact. They are also piloting a service to help identify when recycling has been mis-sorted to correctly inform and eventually fine households that continue to put excessive trash in the recycling streams.

The service also creates value for trucking companies that are part of / contract with the waste management ecosystem by reducing wear and tear on the equipment through a preventative maintenance program. The software helps trucks avoid areas with major road damage, further protecting equipment as well as provides safety updates and fuel metrics. By using AI and long-term data analytics, the company can better predict when a truck is going to be full and preemptively send out an additional truck to take over the route. This saves time, optimizes the process, reduces labor usage, and leads to higher customer satisfaction.  Additionally, having more insight into what is in the incoming recycling stream allows MRFs (Materials Recovery Facilities) and other members of the value chain to also operate more efficiently by reorganizing their sorting lines, re-deploying labor to different sorting streams faster, and rejecting loads with too much garbage before sorting. All these improvements help improve margins for members across the value chain – making recycling more economical. This value creation also does social good in that by making recycling more economical, more companies and cities are likely to invest and use the recycling systems which is overall environmentally positive.

Making money

              Rubicon charges haulers (waste management trucking companies) for the services and app at $190 per truck after a 3-month free trial. Most haulers actually use the app and companies (including their municipal stakeholders) find immense value in the app over those 3 months and typically convert into paying customers. When there are enough haulers in a city that sign up with Rubicon, municipalities are then more incentivized to buy a license for their whole city and their city’s fleet. Contracts are often multi-year, insulating the company from competitors and building customer loyalty while also using the data to further develop and refine the AI technology.


Municipalities have strict contract rules and guidelines they need to follow. Getting in and securing contracts with governments can be a lengthy process and often requires a highly defendable ROI to secure the contract. Thankfully for Rubicon, the company has many contracts that can gives new customers confidence in their product and its value creation potentials.  

Challenges related to the AI include identifying trash from recycling as there is so much variance in what is coming through the stream, managing technology costs as higher accuracy AI involves faster frame capture which is more expensive even if it leads to higher accuracy, and pushback from incumbents who fear job losses/disruption in their industry if the city / company adopts Rubicon.


The company is likely going to expand their range to other municipality-based services such as snow plowing, road salting, street cleaning, water metering, identifying illegal dumping locations, and storm clean-up. They are able to leverage their on the ground route data collection with AI-enabled visual interpretation to provide on-the ground information for cities. Municipalities have limited resources and personnel to execute on all the city needs – often requiring some city services to be delayed for long periods of time leading to unhappy and inconvenienced constituents. Using technology and quickly identifying issues will help municipalities better serve their constituents and more effectively use their limited resources.

What’s next?

The company is doing incredibly well and has both the funds and the technological capabilities to execute on many of their opportunities. As well, they have a well proven and tested ROI that makes customers trust and contract with them. They have very limited customer churn and are a leader in the space. All this leads me to believe that the sky is the limit for this company. They could partner with more advance technology and AI processers to make their products more efficient and better at recognizing objects/distinguishing objects from each other. They could also more aggressively pursue the complimentary opportunities outside of the recycling and waste management to get into “smart city” opportunities and become the go-to “smart city” software provider. Or…. they can move into outer space clean-up!




ASOS: Optimizing customer preference using Machine Learning


FINESSE – Using AI to Tackle Overproduction of Clothes

Student comments on AI in waste/recycling management – smarter city services

  1. Very cool concept! I am very excited about smart cities of the future. I hope that companies like Rubicon will be able to scale once they can demonstrate ROI. Is it possible to use private waste companies as a testing ground?

    1. They do!

  2. Very cool concept! I am still hopeful for a day where all trash is single stream (ie everything goes in one bin, no separating recyclables) and then AI image detection is used at waste processing plants to sort out recyclables en masse with high accuracy. It seems to me that Rubicon is getting a lot of value for using AI in place of otherwise “managerial” decisions, like when to send a new truck out, how often to get it repaired, labor demands and schedules.

  3. Very cool! Maybe I’m misunderstanding their pricing model but it feels like they’re way undercharging for the value being provided. This is a really neat efficiency play for waste management.

    Also weirdly enough Rubicon recently sponsored a competition called Clear Constellation where they sought proposals to solve the problem of space debris. It’s cool to think how they might apply these problems to other “waste management” arenas

Leave a comment