France is using two domestic vendors to make a blunt point: the state wants less dependence on American software in its most sensitive systems. Palantir and the rest of the US enterprise stack should take that seriously.
France has picked a fight that is bigger than one intelligence contract. Prime Minister Sébastien Lecornu's June 16 announcement puts ChapsVision in line to replace Palantir inside the DGSI, France's domestic intelligence service, and gives French civil servants access to an AI assistant powered by Mistral. That is not a routine procurement story. It is the French state telling you where it thinks strategic software power should sit.
The DGSI began using Palantir after the 2015 Paris attacks, when France wanted tools that could help analysts make sense of large, messy pools of security data. Palantir's Gotham platform was built for that kind of work. It is used by defense, intelligence and law enforcement customers, and its whole appeal is that it connects information that would otherwise sit in separate systems.
ChapsVision is trying to take that job away. The French company was founded in 2019 by Olivier Dellenbach, is based in the Paris region, and has built itself through acquisitions including Sinequa, Systran, Deveryware and Bertin IT. According to ChapsVision's public profile and French business reporting cited in company records, its ArgonOS platform is designed to process large volumes of heterogeneous data, the same unglamorous but essential work that made Palantir valuable to governments in the first place.
That is the point. France is not replacing an American contractor with a patriotic slogan. It is betting that a local software company can now handle work that, after 2015, officials did not think a domestic supplier could credibly do. ChapsVision's proposal was selected in the first phase of the DGSI's OTDH programme, the tool for processing heterogeneous data that was designed to move the service away from Palantir. The contract has been reported in France as worth about 40 million euros.
The sovereignty argument is real, but don't let the word make the story vague. For an intelligence agency, the question is not whether a dashboard looks better in French colors. It is whether the most sensitive data tools used by the state should depend on a company incorporated under US law. If you run a European government, that concern is not abstract. It affects who can be pressured, what legal demands can land, and how much leverage Washington has over systems that sit close to national security.
Mistral is becoming part of the state machine
The Mistral part of Lecornu's announcement deserves just as much attention. Giving civil servants a Mistral-powered assistant is a national-scale deployment, not a lab test with a friendly press release attached. France employs roughly 5.7 million people across the public sector, so even partial use would give Mistral something every enterprise AI company wants: a large, visible, demanding customer that can force the product to mature.
The timing helps Mistral too. The Financial Times reported in February that Mistral's annualised revenue run rate had climbed above $400 million, up from about $20 million a year earlier, and that chief executive Arthur Mensch said the company was on track to pass $1 billion by the end of 2026. Those numbers explain why the French state is not treating Mistral as a symbolic European alternative. It is treating it as a supplier that might actually carry weight.
Look at the pattern. ChapsVision gets the sensitive data job. Mistral gets the AI assistant job. Both companies let France say the same thing in two different markets: we don't want every serious software layer to come from the United States. You don't have to be anti-American to understand why that matters. You only have to watch how quickly software, cloud infrastructure and AI models have become part of state power.
For Palantir investors, the honest read is uncomfortable. Palantir's government story has always rested on durability. Once an intelligence or defense agency builds workflows around Gotham, the assumption is that switching becomes painful enough to protect the contract. France is now testing that assumption in the place where Palantir's European reputation was forged after the 2015 attacks.
Frankly, the harder question is not whether European governments want local alternatives. Of course they do. The harder question is whether those alternatives can perform under the pressure of real government work. ChapsVision reported 157 million euros in 2023 revenue and raised 85 million euros in November 2024 to keep expanding, including through the Sinequa acquisition. That is meaningful scale for a French software group. It is still not Palantir scale.
If ChapsVision delivers for the DGSI, the contract becomes a reference case for every European ministry that wants to reduce US dependence without downgrading capability. If it stumbles, Palantir's argument gets stronger: sovereignty is nice, but functioning systems matter more when analysts are under pressure.
France has made its call. Now ChapsVision and Mistral have to prove that sovereign software is not just a procurement preference, but something the state can use every day.
Also read: OpenAI is asking public investors to fund a company that loses three dollars for every dollar it earns • The Hormuz deal Trump announced may not hold and the world economy is not finished with this crisis • STMicroelectronics raises $1.5 billion in convertible bonds as its AI photonics bet reshapes how the market sees the company
