As automotive development accelerates toward the Software-Defined Vehicle, the role of AI at the edge is becoming increasingly critical. In this interview, Stefan Månsby, CEO of Stream Analyze, shares how the company is transforming the way AI models are deployed, monitored, and optimized directly on devices and machines.
At VECS 2026, first-time exhibitor Stream Analyze will showcase its platform for enabling real-time decision intelligence, turning AI models into interactive, queryable systems. By addressing one of the industry’s key challenges – how to continuously evolve AI performance post-deployment – the company is helping OEMs unlock new efficiencies and reduce operational costs. With experience from demanding environments and collaborations with major industry players, Stream Analyze is pushing the boundaries of Edge AI, paving the way for more intelligent, responsive, and scalable automotive systems.
Stream Analyze is joining us as an Exhibitor at VECS 2026 for the first time – please tell us a bit more about your business and area of expertise!
Stream Analyze is a leader in Edge AI, providing a platform to develop, deploy, monitor and operate analytical AI models. We unlock decision intelligence where it matters, on devices, sensors and machines of all sizes, to leverage inference on the edge.
AI is a broadly used tool in the Automotive Industry – how do you see customers currently using AI, and where do you think they could be doing more?
Many AI models, once deployed, suffer from post-deployment stasis. What happens when you need to update the models at scale? Is there a way to reduce the cost of those data stream roundtrips from on-prem, to cloud, to edge and back?
We have pioneered a method of turning your AI Model into a queryable and interactive database. Ask it questions, understand the results and directly optimize the model.
What kind of trends do you think we’ll see more of in the Automotive Industry, and how do you think these will play out long term?
The move toward unified SDV architecture will enable more possibilities for running unique and powerful AI models on the vehicle. For autonomous driving, this means faster evaluation loops for growth in safety for instance – whether it is a perception-based system or an E2E framework using world models for training.
What are you looking forward to discussing with the delegates during this year’s VECS?
We want to understand how OEMs and their key T1/T2suppliers define and currently utilize their edge devices and sensors – and how Stream Analyze’s Edge AI solution might unlock more business cases for better end-customer experiences and lower operational investment along the way.
How does Stream Analyze contribute to the continuous developments of the AI-tools in the automotive industry?
We are already working with major OEMs to “pressure test” our solution in harsh environments, such as closed sites such as construction and mining. Understanding how AI models react in non-connectivity situations, how they work together in federated ML meshes and how ModelOps can bring efficient organizational improvements to an OEM’s processes are just some of the value propositions we are proving every day.
Finally, could you tell us more about the various projects you’re involved in?
Check out our recent effort with Volvo Trucks and Boliden mines, where we turned heavy equipment and their AI models on the machine into queryable databases. There’s more on our website: www.streamanalyze.com