Three Signals From Intertraffic Amsterdam

Martin Blaszczyk, Flasheye, and Torbjörn Halstensen, Opsys Technologies, one of our partners.

We just returned from Intertraffic Amsterdam 2026. Four days, hundreds of conversations, and a lot of coffee. Here's what we actually took away: three deeper shifts happening in static lidar for transportation that are worth paying attention to.

1. The Perception Layer Is Finally Being Treated as a First-Class Problem

For years, the conversation in infrastructure lidar centred on the sensor: range, resolution, form factor, price point. That made sense. Getting the hardware right was the prerequisite. But what we heard clearly at Intertraffic this year is that the industry has reached a new phase, one where the sensor is no longer the limiting factor and perception software is where the real differentiation happens. The questions buyers are asking have shifted. Can the system reliably classify a cyclist versus a scooter in variable lighting? Can it maintain object tracks through occlusion? Can it be reconfigured as traffic patterns evolve without a full redeployment? These are perception questions, and they are driving purchasing decisions in a way they simply weren't two or three years ago. This is exactly the space Flasheye was built for. We work exclusively on the perception side and build the software that turns a raw point cloud into a reliable, actionable understanding of what is happening in a scene. As perception becomes the layer where infrastructure lidar projects succeed or fall short, that focus is what sets us apart.

2. Scalability Requires Rethinking the Power and Processing Equation

One of the quieter conversations at Intertraffic, but one that kept coming up, was about deployment economics at scale. Putting lidar at one intersection is straightforward. Putting it at fifty, or across a rural corridor with no grid infrastructure, is a different challenge entirely.

The bottleneck is rarely the sensor price anymore. It's the power and compute requirements of the full system. Most perception software is built with the assumption of a powered cabinet, a network connection, and a server somewhere. That assumption breaks down fast when you're looking at remote junctions, temporary installations, or simply trying to scale a network without civil engineering costs at every node. The systems that are gaining traction are the ones where perception runs efficiently on low-power edge hardware, making it possible to operate the full stack on solar power alone, no grid connection, no roadside cabinet, no external compute. That changes the economics of where lidar can go and how quickly a network can grow. A deployment that would otherwise require significant infrastructure investment becomes something an operator can roll out incrementally, at far lower cost per site. Scalability stops being a budget problem and starts being a planning conversation.

3. Repeatability is Becoming the Defining Safety Requirement

Infrastructure lidar is moving into applications where the output directly influences safety-critical decisions: signal control, incident detection, speed enforcement, vulnerable road user protection. And in those contexts, a question is starting to be asked more seriously: can you guarantee the system responds the same way every time?

With probabilistic, model-based perception, the honest answer is no. The same scene can produce different classifications depending on factors that are difficult to isolate or explain. Flasheye's perception is built on deterministic algorithms. Given the same input, it will always produce the same output. No variance, no unexplained edge cases, no behavior that drifts over time. That means you can test it systematically, verify it against a defined specification, and stand behind the results. For operators and system integrators who are increasingly being asked to demonstrate safety compliance, repeatability is the property that makes certification possible.

What This Adds Up To

The common thread across all three trends is a shift in where value is actually being created in infrastructure lidar. It's in who can take sensor data and turn it into reliable, maintainable, scalable situational awareness without requiring a research team to operate it. That's the space Flasheye operates in. A focused collaborator that brings the perception layer to partners who build sensors, systems, platforms, and machines.

Want to understand how Flasheye's perception software could work in your system or project? Get in touch.