Chapel Hill (United States) (AFP) – Palantir boss Alex Karp, whom Paris decided in June to do without for its internal intelligence, warned Wednesday that France will “hurt itself” if it believes it is protecting itself with French companies which in reality operate American and Chinese artificial intelligence.
Published on: 02/09/2026 - 23:21
Being satisfied with a French intermediary is an illusory protection, "it's literally the technical equivalent of a 5 to 7 without a condom", he told AFP, in French in the text, on the sidelines of a G20 ministerial meeting in North Carolina.
If France lets its decisions be "guided by fear and by people who don't really understand the issues", "this will hurt France", he added during this meeting organized by the Trump administration to convince its partners not to slow down artificial intelligence with new regulations.
Prime Minister Sébastien Lecornu announced in June, in the name of "strategic autonomy", that he wanted to replace Palantir's data analysis software, used by the DGSI since 2016, with that of the French company ChapsVision. Palantir then assured that its contract, renewed until 2028, remained “fully in force”.
The announcement followed the Trump administration's order for Anthropic to suspend access of "any foreign national" to its two most powerful models, which had alarmed Europeans over their dependence on American technologies.
Without citing ChapsVision, Mr. Karp criticized the idea of entrusting “only” to “French companies” the deployment of American and Chinese “closed models” “without an application layer”.
“The regulatory environment is clearly not helping Europe or France, and you need regulation that really helps,” he added.
He assures that he is not speaking in the interest of American companies, Palantir having "a very small presence" and not "really seeking to expand", but in that of France, "culturally the most important country in Europe" and "economically the second".
“He who loves well punishes well”, justified himself, again in French, this doctor of philosophy graduated in Germany, who runs a company co-founded in 2003 with Peter Thiel, support of Donald Trump.
Asked about the fear that AI arouses, he acknowledged "legitimate concerns", but "we cannot act as if real technical questions were the same thing as alarmism".
© 2026 AFP
The content you are trying to access does not exist or is no longer available.
