Open and reversible: a technological conviction, not just an economic one
Since the group's earliest digital activities, one choice has stood out: favoring open technologies. Why that choice goes beyond simple cost considerations.
The choice of open source technologies is often presented as an economic decision: avoiding costly licenses, cutting operating costs. That is a real benefit, but it is not, for the group, the main reason behind this choice — repeated since its earliest digital activities, through IoT, infrastructure, managed IT services and, today, artificial intelligence.
The deeper reason rests on a conviction: an information system must remain understandable, controllable and evolvable by the people who operate it. A closed architecture can work perfectly well — until the day it has to evolve, be migrated, or simply be understood well enough to explain why it behaves a certain way. That is when the initial technology choice gets paid for, or pays off.
Concretely, favoring open standards limits several forms of dependency: dependency on a single vendor, technological lock-in, loss of control over data, dependency on proprietary architectures. These are not abstract risks — they are very concrete reversibility problems, which tend to surface at the worst possible time: when something has to change fast.
That principle — control rather than depend — applies to artificial intelligence today with the same rigor it once applied to server infrastructure. An AI Native system built on open, documented building blocks stays steerable by the team operating it. An AI system built as a proprietary black box moves that control elsewhere — a trade-off that fifteen years of technical experience have taught the group to avoid by default.