As Europe’s automotive sector faces rising pressure from US and Asian competitors, a Swiss AI firm is positioning itself as a potential ally by helping carmakers optimise how they collect and use data for autonomous driving.
Zurich-based start-up Lightly has launched LightlyEdge, a new tool designed to reduce the volume of redundant data gathered by vehicle cameras and sensors.
The company says the system can identify and retain only the most relevant data for AI training, which could significantly cut storage and bandwidth costs while improving the quality of autonomous vehicle systems.
The challenge of managing vast amounts of data is a growing issue for car manufacturers developing self-driving technologies. As AI models grow in complexity, the data used to train them often includes large amounts of repetition or irrelevant footage—raising both technical and financial hurdles.
The fledgling firm claims its system offers a real-time solution by deploying intelligent models directly onto edge devices, such as cameras and sensors embedded in vehicles.
These models can filter data as it is captured, identifying unusual or critical scenarios, such as pedestrians crossing roads or low-visibility conditions, while ignoring repetitive or non-essential footage.
“LightlyEdge tackles a critical bottleneck in AI development: capturing only the data that truly matters, directly at the source,” the company stated in a press release.
Co-founder Matthias Heller said the technology offers manufacturers a competitive edge by accelerating model development while lowering operational costs.
“By focusing on quality data from uncommon scenarios and hazards, we’re empowering a new era of innovation that will drive the future of mobility,” he said.
The timing of the product launch comes as European carmakers contend with trade tariffs, labour shortages, and aggressive competition from firms like Tesla and China’s BYD. Tesla’s own data strategy—known as Active Learning, which emphasises targeted data collection—has become an industry benchmark.
LightlyEdge builds on the company’s previous product, LightlyOne, which was developed for data centre environments. The new system brings those capabilities directly to vehicles, allowing AI teams to deploy, monitor, and improve models on the road with an accessible user interface and faster iteration cycles.
Founded by former ETH Zurich and Harvard students Matthias Heller and Igor Susmelj, Lightly is one of a growing number of European AI firms aiming to support traditional industries in the transition towards intelligent automation.
The company believes smarter data collection could give European automakers a much-needed edge in the global race to develop autonomous vehicles.
