French deeptech has changed scale. According to Bpifrance, startups in the sector raised 4.1 billion euros in 2025, three times more than in 2019. The ecosystem now includes 2,830 companies. But all is not yet won.
Behind these figures lies a more nuanced reality. A technology can work in the laboratory… but never find its place in industry. Between the first scientific results and adoption by a customer, deeptech startups must still prove the reliability of their technology, adapt it to field constraints and demonstrate its economic relevance. This is the famous “valley of death”: a phase where financing needs are significant, while commercial outlets remain uncertain.
A program to help deeptech startups take the plunge
Nicolas Reydellet knows the subject well, this “delicate milestone to cross”. Since the end of 2025, he has led the EDF Pulse Pilot program, which aims to help startups get through this pivotal moment. The system thus targets technologies from technological maturity levels known as “TRL 4 or 5”.
Its operation is based on a logic of co-development. EDF starts “from an operational need identified in its businesses”, searches for technologies likely to respond to it, then works with the start-ups, laboratories or ETIs concerned to create a demonstrator.
“We are not going to derisk for the sake of derisking. What interests us is to see from the start in the projects an interest for the EDF group, for our businesses, but also for the external market,” explains Nicolas Reydellet.
The financing is at risk but not dilutive. Concretely, this means that the Group does not take any participation in the capital of companies. It invests in the development of the solution via contracts and targets remuneration, in particular through the valorization of the intellectual property created jointly. “The objective is also to give startups a first industrial reference, useful for convincing other clients and investors,” underlines the program director.
This strategy involves a certain risk, and accepting that not all projects will succeed. But therein lies the advantage of being a large Group. EDF can afford to rely on a diversified portfolio, made up of projects of different sizes and maturities. “If there is no risk, it’s because we’re not on the right project,” jokes Nicolas Reydellet. The income generated by successful projects must help finance subsequent ones.
A robot to repair dam pipes
Among the projects supported, EDF Pulse Pilot is, for example, working with EDF Hydro to develop a robot capable of inspecting and repairing the penstocks of its dams. Some of these facilities, built in the 1930s to 1950s, are now affected by corrosion and leaks. Their maintenance also poses safety problems: the pipes can be narrow, difficult to access and contain asbestos or lead.
EDF Pulse Pilot carried out an analysis which identified around a hundred robotic solutions on the market in 16 countries. This analysis concluded that no solution could be used as is and that the market was largely dominated by inspection robots rather than repair robots, particularly in inclined and wet pipes. The Group therefore brought together three French companies to combine their technologies: a rolling robot on the one hand, an articulated arm on the other and finally, an integration system.
The final test stage must still be carried out in the coming months. The solution is therefore not yet commercialized, but “its potential already goes beyond the sole case of dams”. Other professions have expressed interest in the final robot, and discussions are already underway with external manufacturers.
“We built it on the use case of our dams. But we thought of it, as with the other projects, as being scalable,” underlines Nicolas Reydellet. “For partner startups, the challenge is to be able to promote their technology to other clients, beyond this first demonstrator. »
EDF Pulse Pilot currently has more than 20 projects under study or in development, with deployment planned from 2027. Robotics and the electrification of industrial processes are among its priorities, specifies our interlocutor.
Bringing together industrialists and project leaders, the next challenge?
A challenge remains: organizing meetings between manufacturers and technology carriers early enough in their development. “Meetings at this level of maturity are less structured,” notes Nicolas Reydellet.
“The connection has to be made. They need to come see us and we need to go see them. And we need to be able to create the meeting place at this level of development,” he observes.
According to him, the subject goes beyond the sole question of financing. Large groups can play the role of first customer, but also of technological partner, by providing their infrastructures, their teams and their use cases. On condition of agreeing to take their share of the risk… The result is a great reward: not leaving the innovations developed in the laboratory at the door of its factories.

