TUM Startup Uses Optical AI to Sort Up to Ten Tons of Potatoes Per Hour

The founders of Karevo: Johannes von Wittke, Benedikt Keßler and Felix Beck.

(IN BRIEF) Karevo, a spin-off from the Technical University of Munich, has developed an AI-powered potato sorting machine capable of processing up to ten tons of potatoes per hour with 95 percent accuracy. The system was trained on more than 100,000 images and can identify foreign objects and seven types of potato defects, including rot, cracks and wireworm damage. A key feature is its ability to sort unwashed potatoes, which can vary widely in appearance depending on soil, region and storage conditions. Founded by Benedikt Keßler, Johannes von Wittke and Felix Beck, Karevo targets small and family-run farms with a smaller, more affordable and modular system that can be integrated into existing operations.

(PRESS RELEASE) MUNICH, 11-Aug-2026 — /EuropaWire/ — Karevo, a spin-off from the Technical University of Munich, has developed an AI-powered machine that automates potato sorting and can process up to ten tons of potatoes per hour.

The system uses optical recognition supported by a specially trained artificial intelligence model and achieves 95 percent accuracy in identifying defects and unwanted material.

Karevo was founded by Benedikt Keßler, Johannes von Wittke and Felix Beck.

Keßler grew up on a potato farm and developed the idea for the machine after experiencing the physical demands of manual sorting first-hand.

He said sorting potatoes requires workers to stand for hours at sorting tables, inspecting each potato individually while working in noisy and dusty conditions.

While studying mechanical engineering at TUM, Keßler began developing the concept for an automated sorting machine that could reduce this labour-intensive work.

The AI model behind the system was trained using more than 100,000 images.

It can detect foreign objects and identify seven types of defects, including rot, cracks and damage caused by wireworms.

One of the system’s key advantages is that it can also sort unwashed potatoes.

Keßler said other detection methods often struggle with unwashed potatoes because their appearance can vary significantly depending on region, soil type and storage conditions.

Karevo’s AI model can be adapted quickly to the specific potatoes being processed on each farm, allowing the sorting process to reflect real operating conditions.

The company designed its machine especially for small and family-run farms.

Keßler said many potato sorting machines are too expensive, too large or too difficult to maintain for smaller agricultural businesses.

Karevo’s system is smaller and more affordable, while its modular structure is intended to simplify maintenance and make integration into existing farm systems easier.

Keßler said he developed the machine he would have wanted at his parents’ potato farm to remove the hardest part of the manual work.

The technology originated during Keßler’s master’s thesis at TUM.

Together with his friend Johannes von Wittke, an engineering student at the East Bavarian University of Applied Sciences in Regensburg, he began looking for further co-founders.

Through UnternehmerTUM, the Center for Innovation and Entrepreneurship at TUM, the team connected with Felix Beck, who had completed his Executive MBA at TUM.

During the early startup phase, the founders participated in the UnternehmerTUM Incubator.

They founded Karevo in 2024 and initially developed prototypes in the Makerspace, UnternehmerTUM’s high-tech workshop.

Karevo has been successfully on the market since the autumn of 2025.

The company’s machines sort potatoes using an AI-powered optical system trained on more than 100,000 images.

The system transports potatoes into a green sorting module, where they are analysed and sorted before being conveyed into a collection bin.

Karevo is part of TUM’s wider entrepreneurship ecosystem, which supports startup teams through consulting, education, incubation, technical infrastructure and access to market expertise.

The innovation ecosystem around TUM is considered one of Europe’s successful deep-tech hubs, supported by a broad network connecting startups with established companies, experts, investors and public-sector partners.

TUM and UnternehmerTUM provide programmes tailored to different phases of the startup journey.

The TUM Venture Labs support teams across twelve technology fields, offering access to cutting-edge research, technical infrastructure and market knowledge.

Each year, more than 100 companies are founded at TUM, while more than 1,000 startup teams are supported by UnternehmerTUM and the TUM Venture Labs.

UnternehmerTUM also invests through its own venture capital fund and has been named Europe’s best startup centre three times by the Financial Times.

Further information and links

The innovation ecosystem centered around TUM is considered one of the most successful deep-tech hubs in Europe. Its particular strengths lie in its strong, diverse network and highly targeted support. Through initiatives and co-labs, startups collaborate with established companies, experts, investors, and government agencies to drive innovation. TUM and UnternehmerTUM, the Center for Innovation and Entrepreneurship, support startup teams with programs precisely tailored to the individual phases of the startup journey and the teams themselves. Across twelve technology fields, the TUM Venture Labs offer direct access to cutting-edge research, technical infrastructure, and market expertise. Each year, more than 100 companies are founded at TUM, and more than 1,000 startup teams are supported by UnternehmerTUM and the Venture Labs. UnternehmerTUM, which invests through its own venture capital fund, has been named Europe’s best startup center three times by the Financial Times.

Media Contact:

Corporate Communications Center
Linda Schinnenburg
presse@tum.de

SOURCE: Technical University of Munich