In this article, Jérémy Picot-Clémente, EPIC’s Photonics Technologies Program Manager, talks to Luca Verre, Co-Founder and CEO at Prophesee, a French developer of advanced neuromorphic vision systems.
What’s the background to you cofounding Prophesee as CEO?
In 2007, after completing an MSc in Industrial Engineering at the École Centrale de Lyon, in France and an MSc in Physics and Electrical Engineering at the Polytechnic University of Milan, I joined Schneider Electric as a Project manager. For the next five years, I worked in business development and product marketing in Japan, Germany and France.
Although it was a great learning experience, by 2013, I’d become frustrated with working in such a large organization because it was difficult to feel the impact of what I was doing on a daily basis. I ended up deciding to take a sabbatical to pursue an MBA at INSEAD with the idea of eventually resuming my career at Sneider, or better still to work for a smaller company.
While I was at INSEAD, I met Christoph Posch, an Austrian researcher who had been working on neuromorphic technology for the past 25 years. When we met, Christoph was thinking of setting up a company to commercialize his research. I immediately saw the potential of the technology, and as I was also looking for my next career step, in 2014, we decided to set up Prophesee together with Christoph as CTO and me as CEO.
What is neuromorphic technology?
Prophesee specializes in neuromorphic, event-based vision technology that mimics the human visual system. Unlike traditional video cameras that capture images at fixed frame rates, Prophesee's sensors act in the same way as the human eye in that they only record when there’s a change in the scene, i.e. when individual pixels perceive movement or illumination changes.
What are the advantages of event-based sensors compared with conventional frame-based sensors?
Frame-based systems constantly record whether or not anything is happening. This creates a lot of redundant data and both power and storage requirements are high. In contrast, event-based systems only record when something changes, so much less data is recorded, data processing is faster and power and storage requirements are much lower, making them suitable for battery-powered applications.
The other advantage is that event-based capture is considerably faster. With frame-based systems, the entire sensor waits to be read out at a fixed interval, whereas in event-based vision, each pixel operates independently and immediately reports a change when it detects one. This gives event-based systems microsecond latency with near-instant response to changes in the visual scene, making our systems ideal for real-time systems like robotics, autonomous vehicles, or AR/VR, where fast response is needed.
Event-based sensors also provide much higher temporal resolution. In frame-based sensors, temporal resolution is defined and limited by frame rate. However, event-based systems have no fixed frame rate. Each pixel reports changes independently and asynchronously giving event-based systems microsecond-level precision, which is perfect for capturing fast-moving objects or scenes with quick dynamics like drones flying at high speeds.
Additionally, frame-based capture integrates over milliseconds, which can cause blur. In contrast, with event-based systems there’s no integration time because they capture events as they occur, pixel by pixel, which means they provide crystal-clear vision even during fast motion. Another advantage is that because each pixel operates independently and handles light logarithmically, event-based systems have a much higher dynamic range, meaning they can perform well in extreme lighting conditions. Taken together, these features make event-based sensing ideal for battery-powered applications like IoT, wearables and smartphones.
How has the company developed?
In 2015, we built our first prototypes. Since then, we’ve developed a range of the world’s smallest and most power-efficient event-based vision sensors. We also provide camera systems around the sensors and a software development kit that we license to our customers so they can build an entire solution. During this period, we have raised more than €100 million in investment, registered more than 90 international patents, and our workforce has grown to 100+.
What have been the main challenges in growing the company?
The first is investment. We raised the first round from family and friends and after six months we’d raised €750,000. But it was clear from the beginning that bringing such a disruptive technology to market would require a lot of changes in the ecosystem and finding investors willing to take on this risk was difficult. Our first investors were Bosch and Intel Capital who led our A and B rounds. The challenge was not just finding an investor, it was finding an investor who understood the complexity of what we were trying to do, who would be patient and who would be ready for the long journey and support us in our decisions.
The second challenge was getting access to an advanced manufacturing technology node, and one of our major achievements was to convince Sony to give us access to their foundry. Seeing the potential of our technology, Sony agreed to lead our B+ round and to collaborate with us on the development of a new generation of sensor products which we eventually launched in 2021. This was a major milestone for us because not only did we have access to an advanced manufacturing technology that enabled us to prove the technology was economically viable and scalable, but we also became associated with market leader in consumer electronics, which was a significant validation for the company and for the technology.
The third was attracting talent, which for a small company is always difficult. Fortunately, in 2015 we were able to take advantage of the growing excitement in France around startups, and we were able to recruit very skilled talents from top universities in France.
How do you see the future?
We will continue to be a market leader and pioneer in event-based technology. We see a growing need for our technology in the industrial, robotics and AR/VR market segments, as well as defense where we bring huge differentiating value in terms of high speed, real-time detection and tracking of fast-moving objects. Our goal is to keep growing in these spaces and become a top semiconductor company in Europe.
What’s your advice for the next generation of entrepreneurs?
Firstly, raising money will always be a challenge, especially in Europe. Although there is more money available in Europe than previously for deep tech, Europe is still a long way behind China and the US. In 2023, for example, the US invested more than 100 billion in deep tech, China invested 50 billion but Europe invested only 20 billion. The main challenge in Europe is not seed funding or series A, it’s at the D round stage where companies need significant capital, i.e., around $50 million, to scale up.
This is why companies often have to move to the US or license their technology because it's the only way they can survive. So, my advice would be to play globally from the very beginning so you can attract global partners and investors.
Secondly, from day one be very focused on finding the shortest path to revenue because when you generate revenue and you generate profit, you immediately become more independent, less reliant on investors and you gain more control over the future of your company.
Written by Jérémy Picot-Clémente, EPIC’s Photonics Technologies Program Manager.
Back to Features
