Application proof of concepts covering AI, machine learning, and end-user interface
Concept uses predictive maintenance to ensure longer product life cycles
Concept predicts when machine components might need maintenance and proactively addresses it
TOMRA Collection uses data and customer insights to innovate, support recycling, and make a bigger impact on waste reduction.

The growing volume of plastic waste has brought us to a critical juncture globally. At the moment, only 2% of the world's plastic packaging is recycled in a closed-loop system, but with deposit return systems (DRS) for drink container recycling, that can change.
Norwegian brothers Tore and Petter Planke founded TOMRA in 1972 in response to solving a fundamental problem for retailers. They aimed to design the world’s first fully-automated reverse vending machine (RVM)—that accepts empty beverage containers and provides a refund or reward in return.
Fifty years later, TOMRA Collection is feeding the circular economy by merging technology with sustainable impact to lead the resource revolution through its Collection, Recycling, and Food divisions.
TOMRA Collection’s 95,000 reverse vending machines operate in over 60 global markets and transform how it obtains, uses, and reuses materials*. Shoppers visit their local supermarket with empty bottles and cans, insert them into a reverse vending machine, and receive a deposit refund when they’re finished. Returning drink containers for recycling reduces reliance on virgin raw materials and minimizes waste in streets, oceans, and landfills. TOMRA Collection’s reverse vending machines currently capture 1,500 bottles and cans every second—and over 57 billion containers a year.*
*Figures from 2026.
When the EU introduced its directive that member states collect 90% of plastic bottles by 2029, many experts agreed that this target would not be possible without efficient deposit return systems (DRS). Reverse vending machines are a key part of deposit return systems, which see up to 98% of all drink containers returned for recycling.
TOMRA Collection spent decades working on new developments to their machines, releasing cutting-edge RVM technology that opens new opportunities in reverse vending. Its RVM sensors collect data every time a bottle is deposited. With insights into recycling patterns, collection volumes, and material composition, TOMRA Collection can make data-driven decisions to improve its machines, from design and experimentation to consumer use.
“Our ambition is to have machines that are always available and easy to operate,” says Marius Sommerseth, Head of Digital Products at TOMRA Collection. “TOMRA Collection started connecting machines 35 years ago, and by using the data we have built up over several years, we’ve gained a lot of insight that has enabled us to offer new models and services to meet customer needs.”
TOMRA Collection’s goal is to collect 500 billion containers per year, which will create a petabyte-scale of data. With an innovative cloud partner like Google, TOMRA Collection has the flexibility and scalability to constantly improve its services toward its customers.
Using Google Cloud technology, TOMRA Collection built the Smart Service concept and analyzed data concurrently. After setting up a framework and data model in BigQuery, TOMRA Collection used Firebase, which simplifies app development, to showcase the solution from a demo perspective. Finally, integrating their data with Pub/Sub enabled TOMRA Collection to pull data in real time.
Google Cloud made our prototyping process significantly faster, which gave us more time for in-depth analysis and the ability to showcase proofs of concepts to our customers quickly. The combined insights from BigQuery, Gemini, and direct feedback from our customers are helping us create even better products.
Mariusz Szerepko
Data Scientist, TOMRA Collection
We are always working to reduce waste and implement sustainability in our product design operations. First, by doing more with less materials and energy. And second, by ensuring our products’ components last longer, and that as many components as possible are reusable and recyclable at the end of their life cycle.
Mickaël Fontaine
VP, Head of Technology, TOMRA Collection
Besides its goal to hit net zero by 2050, TOMRA Collection has other ambitious aims for 2030, like reducing operational transport by 80% and ensuring 50% of its products can be reused.
“We are always working to reduce waste and implement sustainability in our product design operations,” says Mickaël Fontaine, VP, Head of Technology, TOMRA Collection. “First, by doing more with less materials and energy. And second, by ensuring our products’ components last longer, and that as many components as possible are reusable and recyclable at the end of their life cycle.”
For example, TOMRA Collection's Smart Service concept uses predictive maintenance to ensure longer product life cycles. And with Google Cloud technology, the company processes huge volumes of data from its reverse vending machines, predicting when machine components might need maintenance and proactively addressing it.
This increases machine uptime and saves customers the cost of unnecessary or too-late repairs.
TOMRA Collection’s tech-driven approach to reducing waste and improving recycling efficiency fuels its ambition to complete a circular economy and process 500 billion containers annually. And it’s enabling millions of people around the world to adopt practical, habitual sustainability in their everyday lives.
“Customer centricity is key for everything we do,” says Sommerseth. “We listen to our customers to continuously improve our product offering. We want their voices to be heard, because we want to do this together and give them the maximum possible value.” For example, when TOMRA Collection previewed the Smart Service concept at the EuroShop retail trade fair, it didn’t just showcase a proof of concept—it provided the opportunity for customers at the event a chance to immediately give feedback.
With Google Cloud, we can prototype, learn, and iterate quickly. It gives our customers the opportunity to focus on their customers, and gives us a chance to innovate more tailored solutions to help the world recycle. It’s helping us make a positive difference for communities, businesses, and ultimately, the planet.
Marius Sommerseth
Head of Digital Products, TOMRA Collection
The Smart Service concept is taking a data-driven approach and using machine learning to offer proactive notifications about actions required at the machine, and information on cleaning performance and machine usage to better plan day-to-day operations. It also provides comprehensive updates when technicians carry out services, whether remotely or in-person. This offers customers more transparency, avoids unplanned repairs and ensures maximum machine up-time.
“With Google Cloud, we can prototype, learn, and iterate quickly,” says Sommerseth. “It gives our customers the opportunity to focus on their customers, and gives us a chance to innovate more tailored solutions to help the world recycle. It’s helping us make a positive difference for communities, businesses, and ultimately, the planet.”