Summary

• Sortrace has raised SEK8 million in a round led by Almi Invest, with participation from Gorilla Capital, Antler, and private investors.
• Its system uses cameras mounted on collection vehicles and AI-based image analysis to classify waste and identify sorting errors by bin and route.
• The company reports five paying municipal customers in Sweden and plans further commercial expansion and software development.

Swedish waste technology company Sortrace has raised SEK8 million to expand a monitoring system that uses cameras and artificial intelligence to analyse household waste during collection, giving municipalities more detailed information about recycling performance and incorrectly sorted materials.

The investment was led by Almi Invest, with participation from Gorilla Capital, Antler, and private investors. The Stockholm-based company intends to use the funding to expand its customer base and further develop software that converts images captured by collection vehicles into information about the composition of individual waste streams.

Sortrace has developed equipment that can be installed on refuse collection trucks, recording waste as containers are emptied. Its computer vision technology examines the images to identify different material categories and potential contamination, linking those observations to particular bins and collection routes.

The company was founded by Rahel Surshi and Tomas Kjellstrand and began operating its first system in October 2025. It reports five paying customers among Swedish municipalities and municipal waste companies, with discussions under way about additional installations.

Collecting information alongside waste

Municipal waste operators manage large quantities of material, but the information collected during normal operations often provides an incomplete picture of what households discard. Vehicle weights and processing facility records can establish how much material enters the system, while revealing less about where incorrectly sorted waste originated.

That distinction affects efforts to improve recycling. When inappropriate materials enter a collection stream, they can contaminate other waste, increase the work required at processing facilities, and reduce the quantity suitable for recovery.

Local authorities may investigate contamination through manual inspections, sampling exercises, or information gathered at sorting plants. These approaches can identify recurring problems, although the results may represent only a proportion of collections or provide limited information about individual locations.

Sortrace is attempting to obtain more regular observations from the collection process itself. Camera hardware installed on vehicles captures images during bin emptying, while software classifies the visible material and generates records that operators can examine through dashboards.

The resulting information can be associated with routes and individual collection points, enabling municipalities to identify locations where particular sorting errors occur repeatedly. A council could use those observations to investigate why food waste is entering residual rubbish, for example, or whether recyclable packaging is being placed in an unsuitable collection stream.

Targeted household information, changes to container arrangements, or improvements in collection instructions could then be assessed against subsequent observations. Instead of assuming that an intervention has worked because a campaign has been delivered, operators would have a more detailed record of what is entering their collection vehicles.

From image classification to operational decisions

The technology relies on computer vision, in which software interprets visual information to identify objects or material characteristics. In waste collection, the task is complicated by overlapping objects, damaged packaging, variable lighting, and materials that are difficult to distinguish from their appearance alone.

A camera may capture only the visible portion of a container’s contents, while different materials can look similar after being crushed, mixed, or contaminated. The accuracy of automated classifications therefore needs to be assessed against the conditions encountered during ordinary collection activity.

Sortrace’s software is designed to identify material categories and sorting errors rather than require collection workers to examine every load manually. The company’s dashboards allow operators to view patterns across routes and over time, providing a basis for deciding where further investigation or intervention may be worthwhile.

Such information can support decisions about how resources are allocated, although automated observations need careful interpretation. A classification error could indicate a problem at a collection point that does not exist, while an undetected contaminant could leave an operator with an incomplete assessment of waste quality.

Validation against physical samples and consistent measurement methods would help establish how reliably the system performs as it is deployed across different vehicles and municipalities. Local waste collection practices, container designs, and the materials accepted for recycling can also differ, requiring classification and reporting systems to reflect the arrangements used by each customer.

The value of the resulting data will depend on whether local authorities can use it to change operations. More detailed measurements may identify opportunities to improve sorting, but they cannot independently provide the processing facilities, household participation, or markets for recovered materials needed to achieve higher recycling rates.

Municipal deployments provide early evidence

Sortrace is already working with Swedish municipalities and waste management organisations. Its website identifies relationships involving Östersund, Solna, and other participants in the waste technology and research sector, while the company says its paying municipal customer base has reached five.

The Östersund municipality has described using the technology to improve its understanding of waste sorting and support more targeted action. Its experience provides an operational example of how collection data can inform local decisions, although the company has not published an independently verified comparative study establishing particular reductions in contamination or increases in recycling rates.

Funding from Almi Invest and the other participating investors will support further commercial development. For Sortrace, expansion involves more than distributing additional cameras because each customer must be able to integrate observations into its existing collection and management processes.

Municipalities purchasing the service will also need to consider installation and maintenance costs, data availability, staff training, and the resources required to act on the information generated. The value of more detailed observations may differ between operators depending on the scale of their collection networks and the problems they are attempting to address.

Privacy within collection operations

Linking waste observations to individual bins creates data protection considerations because collection records may be associated with specific households or addresses. Even where the purpose is to classify materials rather than identify individuals, the collection and use of location-linked information require appropriate safeguards.

Gorilla Capital says Sortrace anonymises images on the collection vehicle and stores information within the European Union. The company’s website also describes encrypted data transmission and processing through European cloud infrastructure.

Those measures may reduce some privacy risks, but municipalities must establish how the resulting records are used, who can access them, and whether any information can be connected to identifiable households. An observed sorting error does not necessarily establish which individual placed the material in a container or the circumstances in which it was discarded.

Such distinctions become particularly important if the information is used for direct household feedback or enforcement. Automated detection can help direct attention towards a problem, while decisions affecting residents require evidence and procedures appropriate to the action being considered.

Expanding beyond initial installations

Sortrace plans to use the new capital to develop its offering and reach more customers, including organisations outside its initial municipal market. Its technology could potentially be applied at additional stages of waste management where visual information helps operators assess material flows and contamination.

Connecting observations from collection with information gathered at sorting and processing facilities could eventually provide a more complete picture of what happens to material after it enters the waste system. Operators would still need consistent classifications and records to compare results from different points in that process.

The company’s next commercial phase will involve extending its monitoring equipment and software across a wider range of collection arrangements. Its five paying municipal customers provide an initial operating base from which to assess classification performance, implementation costs, and the effect of better information on recycling decisions.