Investing in Annotell: Making safe perception possible for self-driving cars


Annotell banner

We are excited to announce that together with Metaplanet (Jaan Tallinn) we have co-led Annotell’s $24 million Series A round along with participation from their existing investors Stena Sessan and Ernström Kapital. You can read more about the announcement from Techcrunch.

Safety is a paramount issue to tackle before autonomous driving (AD) vehicles are accepted on public roads. Previously a few fatal accidents involving self-driving cars from the likes of Tesla and Uber have made major headlines. This compares to currently an estimated 1.3 million people dying in traffic accidents every year. It is expected that autonomy will lower the number of road accidents significantly, but the accepted safety standard for autonomous vehicles needs to aim much higher than that required from human drivers.

Solving the complexity of safety starts from developing safe perception. However, to get there the AD developers need to collect massive amounts of sensor data as they face the challenge of trying to define safe behavior for all the millions of possible edge cases. Furthermore, they need to consider development speed and cost with ensuring very high quality of annotations where humans train algorithms to classify objects like pedestrians and cars correctly in the collected data.

Annotell’s software platform makes it possible to develop and validate safe perception systems for both AD and advanced driver assistance systems (ADAS) by making sure it is done with the right data. Essentially Annotell provides the “vision test” for the cars to approve their perception systems. They combine software with deep expertise to quantify perception system performance and dataset quality to meet the safety requirements of self-driving cars. This will allow developers to reduce their cost and shorten production time without compromising on safety.

Currently the company has a clear focus on safe perception for self-driving cars given this difficult use case is being industrialized in the automotive side. Yet in the long term their solution could also be applied to other use cases where computer vision may be used in safety critical applications.

Founded as recently as 2018 in Gothenburg, Sweden’s largest automotive cluster, by Daniel Langkilde and Oscar Petersson, the Annotell team has already 70 members and another office in Munich, Germany. They are growing fast and have attracted talent from Mapillary, leading automotive technology companies and OEMs, among others. The company has a global client base with several prominent automotive OEMs and suppliers already as its customers including Zenseact (subsidiary of Volvo Cars) and Magna.

Annotell provides the tools to prepare safe perception for autonomous mobility, and this is why we at NordicNinja were thrilled to join Jaan Tallinn’s Metaplanet to co-lead the latest investment round. We’re excited to help the team scale globally, and build Annotell as a standard for developing and validating safe perception systems for the rapidly growing number of increasingly automated cars.

To learn more about Annotell, please visit or see our portfolio page

Annotell is on a mission to make safe perception for autonomous mobility possible. By combining software with deep expertise to quantify perception system performance and dataset quality to meet the very high requirements that come with developing and validating safe perception systems for self-driving cars, Annotell allows carmakers to reduce their development cost and shorten their production timeline without compromising on safety. Annotell has a global client base and closed deals with several prominent AD companies. Annotell has offices in Gothenburg, Sweden as well as Munich, and Germany.

What’s new in our world