France has taken a highly symbolic step towards digital autonomy, announcing that its domestic intelligence service, the DGSI, will sever its contract with the American data-analytics firm Palantir and migrate to a platform built by the French startup ChapsVision. The decision, communicated by Prime Minister Sébastien Lecornu, is driven by a conviction that strategic dependence on foreign AI tools is no longer tenable. “We cannot rely on instruments developed by foreign powers,” Lecornu declared, framing the move as essential to safeguarding national sovereignty in an era when data has become the new gold. Yet the transition will be far from swift. Roland Lescure, the French economy minister, cautioned that deploying the ChapsVision solution would take “several months, probably in the course of 2027,” owing to a host of technical complexities that must first be resolved.

Viewed from Washington, the French pivot is the latest ripple in a transatlantic rift over who controls the most advanced AI systems. The United States has recently moved to block access for foreign entities to its most powerful models, a decision that has hit European allies particularly hard. Almost simultaneously, a Swiss commercial court issued a ruling that underscored the same logic: artificial intelligence is not a neutral, globally traded commodity but a strategic asset subject to the political priorities of the state that owns it. The case involved Palantir, the surveillance and data-analysis behemoth co-founded by Peter Thiel, a prominent libertarian and early backer of Donald Trump. The company’s chief executive, Alex Karp, has further rattled European sensibilities by publishing an unusually ideological manifesto, reinforcing perceptions that Palantir’s technology comes with a political agenda.

In Rome, the Italian government is watching these developments with quiet resolve. Officials have registered the French decision without surprise and are adopting a similar, if less dramatic, posture: no public attacks, no abrupt ruptures, but a firm halt to any new strategic contracts with Palantir. The underlying fear, shared across European capitals, is that entrusting sensitive institutional and citizen data to a firm so closely associated with a former and possibly future US president represents an unacceptable vulnerability. Italy’s approach mirrors a broader European reassessment of dependencies on American tech platforms, particularly those that operate at the intersection of national security and artificial intelligence.

Germany and France, meanwhile, are accelerating their own bilateral efforts to build a sovereign AI capability. Against the backdrop of Washington’s restrictions, the French national research institute INRIA and the German Research Centre for Artificial Intelligence (DFKI) are expected to announce the first steps towards a joint research and development centre at the VivaTech fair in Paris. The initiative, long discussed but now given fresh urgency, aims to pool European expertise and reduce reliance on American models. Berlin and Paris view the collaboration as a necessary counterweight, not only to Palantir’s dominance in data analytics but also to the broader concentration of AI power in a handful of US firms.

Yet the path to genuine European autonomy in artificial intelligence remains strewn with obstacles. ChapsVision’s platform will require months of integration and testing before it can match the capabilities of the Palantir system it is meant to replace, and the Franco-German research centre is still in its infancy. Analysts in London note that while the political will to decouple from American technology has never been stronger, Europe’s fragmented digital market and chronic underinvestment in AI research could slow progress. The coming months will test whether the continent can translate its newfound wariness of Silicon Valley’s reach into a durable, home-grown alternative — or whether the rhetoric of sovereignty will once again outrun the reality of implementation.