Do French robotics startups share the same research partners?

Take one line of our record. In 2025 the EU funded STARLight, “300mm Silicon Technology for Applications Relying on Light with Photonics Devices”. Among its participants: SteerLight, one of the 89 French robotics companies we follow, and CEA-LIST and CNRS. Its objective, in the register’s own words, opens the panel. That is all a project record says: these organisations were partners in this project. Not who worked with whom inside it, not who paid whom, not whether a product came out of it. Every line in the next screens is a record like this one, and we read them for what they are — a map of who was listed together.

32 of the 89 companies appear in a publicly funded project beside a research organisation. Two names come up most: CEA-LIST, the CEA’s laboratory for digital systems, beside 12 companies, and the CNRS, beside 17. Seven companies share a project with both, five with CEA-LIST alone, ten with the CNRS alone — and ten with neither. Ten of 32 is not a small remainder. The two big names cover most of the companies we can see in a project, not all of them, and the next screen is about the rest.

Across those 32 companies we count 158 distinct research organisations, in France and abroad. Six companies reach one organisation, five two, 21 three or more; the median company reaches four. After the two big names come Delft University of Technology, Virtual Vehicle Research GmbH and INRAE. Breadth here is breadth of co-participation. One large consortium lists a company beside dozens of organisations at once, and every one becomes a dot; a startup with a single big EU project can look better connected than one with three small French ones. The matrix shows where the record puts a company; it does not rank the companies.

Two of the projects, in their own words, are on the panel: photonixFAB, with Aryballe Technologies, and LASTSTEP, with eLichens. An objective is a statement of intended work, written to win the grant; it names a problem and a plan. That is what the record lets us read: which companies were listed with which research organisations, and what those teams set out to do. What it cannot tell us is what came of it — a result, a product, a person moving between them — and we stop here rather than guess. The organisations named on these screens are in the directory, each with its projects.